DETAILED ACTION
Notice of AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
Applicant’s Amendment and remarks dated 1/26/2026 have been considered. Claims 1-20 are pending.
Claim Objections. The objections to claims 1-2, 11, and 16 are withdrawn in view of Applicant’s amendments to such claims.
35 U.S.C. 112(b) Rejections. The rejections to claims 3-6, 8, 12-15, and 17-20 are withdrawn in view of Applicant’s amendments to such claims.
Response to Arguments
On pages 12-13 of Applicant’s 1/26/2026 Amendment and remarks, with respect to the rejections under 35 U.S.C. 101, Applicant argues that under Step 2B, that the claims relate to “improving the technology of energy load prediction.”
The examiner respectfully disagrees. MPEP 2106.05(a) explains that “if the specification explicitly sets forth an improvement but in a conclusory manner (i.e., a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology.” Here, Applicant cites to paras. 0003-0005 as allegedly setting forth the improvement, but after review, the examiner finds that such paragraphs merely explain that there is a need for better load prediction techniques, and then conclusorily states that the present disclose provides a load prediction method and apparatus that is allegedly an improvement. After reviewing the disclosure, the claimed invention appears to merely apply generic neural network training and verification techniques to particular data types, and to implement a particular “hybrid particle swarm optimization algorithm”, without providing any technological improvements to computer functionality, neural network technology, or other technical field. While Applicant asserts that there is an improvement to “technology of energy load prediction”, predicting an energy load is a mental process, so any improvement from the claimed invention is to the mental processes themselves, and not to any actual technology or technical field. Patents that do no more than claim the application of generic machine learning to new data environments, without disclosing improvements to the machine learning models to be applied, are patent ineligible under 35 U.S.C. 101.
On pages 13-14 of Applicant’s 1/26/2026 Amendment and remarks, with respect to the rejections under 35 U.S.C. 103, Applicant argues the following with respect to the SINGH and BOETTCHER references.
PNG
media_image1.png
232
698
media_image1.png
Greyscale
PNG
media_image2.png
202
694
media_image2.png
Greyscale
The examiner respectfully disagrees. SINGH teaches actually verifying the neural network using actual data from 368 days in 2011 (corresponding to recited “n represents a load sampled number”) where the results are measured using MAPE to actually verify the error between predicted and actual data. While SINGH does not teach verifying specifically using thermal load data, BOETTCHER teaches thermal load data, and it would have been obvious to one of skill in the art to supplement the training and verification data of SINGH with the thermal load data of BOETTCHER if available.
On page 14 of Applicant’s 1/26/2026 Amendment and remarks, with respect to the rejections of dependent claims under 35 U.S.C. 103, Applicant argues that such dependent claims should be allowed for the same reasons argued with respect to claim 1.
The examiner respectfully disagrees for the same reasons explained above with respect to claim 1.
Claim Objections
Claims 1 and 7 are objected to because of the following informalities:
In claim 1, line 4, “actually verifying a neural network model” should read “actually verifying the neural network model” because the recited neural network model was previously introduced in line 3.
In claim 7, line 6, “actually verifying a neural network model” should read “actually verifying the neural network model” because the recited neural network model was previously introduced in line 5.
Appropriate correction is required.
Claim Rejections - 35 USC § 101
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding Step 1 of the Alice/Mayo framework, Claims 1-6 are directed to a method (a process), Claims 7-8 are directed to an apparatus (a machine), and Claims 10 and 16-20 are directed to an electronic device (a machine), which each fall within one of the four statutory categories of inventions.
Claims 9 and 11-15 are directed to a “non-volatile memory.” After further consideration, the examiner finds that a “non-volatile memory” would be understood by one of ordinary skill in the art not to cover transitory forms of signal transmission, such as a carrier wave or compact disc, as explained by MPEP 2106.03 II. The examiner notes that a “memory” is appreciably narrower than a “storage medium” and that one of ordinary skill in the art would understand that “non-volatile memory” is a term well-known in the art that refers to a type of computer memory that retains stored information even after power is removed, which excludes transitory forms of signal transmission (which necessarily require power for such transmission). Therefore, claims 9 and 11-15 are directed to an article of manufacture under Step 1.
Regarding Claim 1
Step 2A, prong 1 (Is the claim directed to a law of nature, a natural phenomenon or an abstract idea).
Claim 1 recites the following mental processes, that in each case under the broadest reasonable interpretation, covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components (e.g., “computing device”, and “neural network model”).
An energy load prediction method, for energy supply users (under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally predict the energy load required for energy supply users in a particular building or geographic area)
predicting an energy load (under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally predict the energy load, e.g., the energy load required for energy supply users in a particular building or geographic area, and/or the energy load output by a particular power generating device)
The examiner further notes that the recited “hybrid particle swarm optimization algorithm” is a mathematical concept.
Step 2A, prong 2 (Does the claim recite additional elements that integrate the judicial exception into a practical application?).
The judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements (e.g., “computing device”, and “neural network model”) which are recited at a high-level of generality such that they amount to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
Regarding the “based on a neural network, executed by a computing device” limitation, such limitations are recited at a high-level of generality and amount to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional elements of a computing device and a neural network. These additional elements are recited at a high-level of generality and amount to no more than mere instructions to apply the exception using generic computer components (computing device and a neural network). Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Regarding the “receiving a time period to be predicted” limitation, such additional element of a data gathering step is recited at a high level of generality and amounts to extra-solution activity of receiving data, i.e. pre-solution activity of gathering data for use in the claimed process (see MPEP 2106.05(g)).
Regarding the “training and actually verifying a neural network model ..., wherein training and actually verifying a neural network model ... comprises training the neural network model based on a hybrid particle swarm optimization algorithm, and actually verifying the neural network model based on n true thermal load values and n thermal load predicted values, of a predetermined number of days, through measurement indexes of experimental results which indexes include MAPE, RMSE and DR, n represents a thermal load sampled number, without considering influences of weather, temperature and start-stop state on load” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional element of training and verifying a neural network model to implement a particular mathematical concept using particular data types. This additional element is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (training and verifying of a neural network model). Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)). Moreover, the particular items of training and verifying data merely describe a data type, and therefore such limitation amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use (the field related to energy load parameters). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not integrate a judicial exception into a practical application.
Regarding the “inputting the time period into the neural network model for ... ” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional element of a neural network model. This additional element is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (a neural network model). Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Regarding the “using the neural network model to output an energy load value in the time period so as to reduce a prediction deviation and improve an accuracy of energy load prediction” limitation, such limitation amounts to extra solution activity because it is a mere nominal or tangential addition to the claim, amounting to mere data output (see MPEP 2106.05(g)).
Step 2B (Does the claim recite additional elements that amount to significantly more than the judicial exception?)
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements (e.g., “computing device”, and “neural network model”) are recited at a high-level of generality such that they amount to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
Regarding the “based on a neural network, executed by a computing device” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitation merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)).
Regarding the “receiving a time period to be predicted” limitation, as discussed above, the additional element of a data gathering step is recited at a high level of generality and amounts to extra-solution activity of receiving data, i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Regarding the “training and actually verifying a neural network model ..., wherein training and actually verifying a neural network model ... comprises training the neural network model based on a hybrid particle swarm optimization algorithm, and actually verifying the neural network model based on n true thermal load values and n thermal load predicted values, of a predetermined number of days, through measurement indexes of experimental results which indexes include MAPE, RMSE and DR, n represents a thermal load sampled number, without considering influences of weather, temperature and start-stop state on load” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitation merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)). Moreover, such limitation amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use as explained above, which does not amount to significantly more than the judicial exception. MPEP 2106.05(h).
Regarding the “inputting the time period into the neural network model for ... ” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitation merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)).
Regarding the “using the neural network model to output an energy load value in the time period so as to reduce a prediction deviation and improve an accuracy of energy load prediction” limitation, this limitation amounts to extra solution activity because it is a mere nominal or tangential addition to the claim, amounting to mere data output (see MPEP 2106.05(g)). The courts have similarly found limitations directed to displaying a result, recited at a high level of generality, to be well-understood, routine, and conventional. See (MPEP 2106.05(d)(II), "presenting offers and gathering statistics.", “determining an estimated outcome and setting a price”)
Regarding Claim 2
Step 2A, Prong 2
Regarding the “wherein before the step of inputting the time period into the neural network model for predicting the energy load, the load prediction method further comprises: acquiring the neural network model from a third party” limitation, such additional element of a data gathering step is recited at a high level of generality and amounts to extra-solution activity of receiving data, i.e. pre-solution activity of gathering data for use in the claimed process (see MPEP 2106.05(g)).
Regarding the “obtaining the neural network model by training using sample data” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional element of training a neural network model using sample data. This additional element is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (training a neural network model using sample data). Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Step 2B
Regarding the “wherein before the step of inputting the time period into the neural network model for predicting the energy load, the load prediction method further comprises: acquiring the neural network model from a third party” limitation, as discussed above, the additional element of a data gathering step is recited at a high level of generality and amounts to extra-solution activity of receiving data, i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Regarding the “obtaining the neural network model by training using sample data” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitation merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)).
Regarding Claim 3
Step 2A, Prong 1
wherein the step of obtaining the neural network model by the training using sample data comprises: determining a topology of an initial model, wherein the topology comprises: an input layer, a hidden layer, and an output layer; (under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally decide on the topology for a neural network, such as a 3-layer neural network with an input layer, a single hidden layer, and an output layer)
encoding parameters of the initial model to obtain an initial particle population, wherein the parameters comprise: a center parameter of a radial basis function, a variance parameter, a weight parameter of the hidden layer, and a weight parameter of the output layer, and each particle has at least one parameter; (under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally determine numerical values in an appropriate format to act as the center parameter, variance parameter, and weight parameters for each particle)
decoding the initial particle population to obtain initial parameters of the initial model; (under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally convert particle parameters into a format appropriate for the initial model)
Step 2A, Prong 2
Regarding the “assigning the initial parameters to the initial model to obtain an RBF network model” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional element of initializing a neural network model using initial parameters. This additional element is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (initializing a neural network model using initial parameters). Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Regarding the “optimizing the RBF network model by using a training sample and a test sample” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional element of optimizing a neural network model using training/test data. This additional element is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (optimizing a neural network model using training/test data). Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Step 2B
Regarding the “assigning the initial parameters to the initial model to obtain an RBF network model” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitation merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)).
Regarding the “optimizing the RBF network model by using a training sample and a test sample” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitation merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)).
Regarding Claim 4
Step 2A, Prong 1
selecting a norm of an error matrix consisting of the test value and the expected value as a fitness value; (under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally find a norm of the error matrix and mentally select it to be used as a fitness value)
updating particles in the initial particle population by using the fitness value. (under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally change the mathematical representation of the particles using the fitness value)
Step 2A, Prong 2
Regarding the “wherein the step of optimizing the RBF network model by using the training sample and the test sample comprises: inputting the training sample and the test sample into the RBF network model respectively to obtain a test value and an expected value” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional element of merely inputting a value into a neural network to receive a result. This additional element is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (inputting a value into a neural network to receive a result). Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Step 2B
Regarding the “wherein the step of optimizing the RBF network model by using the training sample and the test sample comprises: inputting the training sample and the test sample into the RBF network model respectively to obtain a test value and an expected value” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitation merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)).
Regarding Claim 5
Step 2A, Prong 1
wherein the step of updating the particles in the initial particle population by using the fitness value comprises: updating velocities and positions of the particles in the initial particle population; (under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally update the values for the velocities and positions of the particles with respect to the mathematical formulas representing such velocities and positions of the particles)
updating an individual extremum of the particles in the initial particle population by using the fitness value, and updating a population extremum of the particles in the initial particle population by using the fitness value; (under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally update the values for the local and global extremums of the particles with respect to the mathematical formulas representing such local and global extremums of the particles in view of the fitness value)
mutating the particles in the initial particle population, and updating the particles when a fitness value of a new particle is better than a fitness value of an old particle. (under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally update parameters for the particles when a fitness value of a new particle is better than a fitness value of an old particle)
Regarding Step 2A, Prong 2, the claim does not include any additional elements that integrate the judicial exception into a practical application and regarding Step 2B, there are no additional elements recited that amount to significantly more than the judicial exception.
Regarding Claim 6
Step 2A, Prong 1
wherein the step of updating the velocities and the positions of the particles in the initial particle population comprises: iteratively updating the velocities and the positions of the particles in the initial particle population by using the following formulas:
PNG
media_image3.png
348
688
media_image3.png
Greyscale
(under the broadest reasonable interpretation, this limitation can be performed mentally by a human, for example, a human can mentally update parameters for the particles using the claimed mathematical formulas; this is also a mathematical calculation, which is another type of abstract idea)
Regarding Step 2A, Prong 2, the claim does not include any additional elements that integrate the judicial exception into a practical application and regarding Step 2B, there are no additional elements recited that amount to significantly more than the judicial exception.
Regarding Claim 7
Step 2A, Prong 1
Claim 7 recites an apparatus that corresponds to the method of claim 1, and therefore the analysis under Step 2A, Prong 1 with respect to claim 1 also applies to this claim 7. While claim 7 recites additional generic computing components (“neural network”, “processor”, “neural network model”, and generic “modules”), such additional generic computing components do not change the analysis under Step 2A, Prong 1.
Step 2A, Prong 2
Claim 7 recites an apparatus that corresponds to the method of claim 1, and therefore the analysis under Step 2A, Prong 2 with respect to claim 1 also applies to this claim 7. While claim 7 recites additional generic computing components (“neural network”, “processor”, “neural network model”, and generic “modules”), such additional generic computing components do not change the analysis under Step 2A, Prong 2.
Step 2B
Claim 7 recites an apparatus that corresponds to the method of claim 1, and therefore the analysis under Step 2B with respect to claim 1 also applies to this claim 7. While claim 7 recites additional generic computing components (“neural network”, “processor”, “neural network model”, and generic “modules”), such additional generic computing components do not change the analysis under Step 2B.
Claim 8 depends from claim 7 and claims an apparatus that corresponds to the method of claim 3, and is therefore rejected for the same reasons explained above with respect to claims 3 and 7.
Claim 9 recites a “non-volatile memory” and further recites generic computing components (computer program) that perform the method of claim 1, and is therefore rejected for the same reasons explained above with respect to claim 1. The addition of generic computer components (non-volatile memory, computer program) does not change the analysis under Step 2A, Prong 2, and Step 2B as previously explained with respect to claim 1.
Claim 10 recites an electronic device and further recites generic computing components (memory, processor, computer program) that performs the method of claim 1, and is therefore rejected for the same reasons explained above with respect to claim 1. The addition of generic computer components (non-volatile memory, processor, computer program) does not change the analysis under Step 2A, Prong 2, and Step 2B as previously explained with respect to claim 1.
Claim 11 depends from claim 9 and claims a memory that corresponds to the method of claim 2, and is therefore rejected for the same reasons explained above with respect to claims 2 and 9.
Claim 12 depends from claim 11 and claims a memory that corresponds to the method of claim 3, and is therefore rejected for the same reasons explained above with respect to claims 3 and 11.
Claim 13 depends from claim 12 and claims a memory that corresponds to the method of claim 4, and is therefore rejected for the same reasons explained above with respect to claims 4 and 12.
Claim 14 depends from claim 13 and claims a memory that corresponds to the method of claim 5, and is therefore rejected for the same reasons explained above with respect to claims 5 and 13.
Claim 15 depends from claim 14 and claims a memory that corresponds to the method of claim 6, and is therefore rejected for the same reasons explained above with respect to claims 6 and 14.
Claim 16 depends from claim 10 and claims an electronic device that corresponds to the method of claim 2, and is therefore rejected for the same reasons explained above with respect to claims 2 and 10.
Claim 17 depends from claim 16 and claims an electronic device that corresponds to the method of claim 3, and is therefore rejected for the same reasons explained above with respect to claims 3 and 16.
Claim 18 depends from claim 17 and claims an electronic device that corresponds to the method of claim 4, and is therefore rejected for the same reasons explained above with respect to claims 4 and 17.
Claim 19 depends from claim 18 and claims an electronic device that corresponds to the method of claim 5, and is therefore rejected for the same reasons explained above with respect to claims 5 and 18.
Claim 20 depends from claim 19 and claims an electronic device that corresponds to the method of claim 6, and is therefore rejected for the same reasons explained above with respect to claims 6 and 19.
Claim Rejections - 35 USC § 103
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 7, and 9-10 are rejected under 35 U.S.C. 103 as being unpatentable over Singh, Navneet Kumar, et al. "PSO optimized radial basis function neural network based electric load forecasting model." 2014 Australasian Universities Power Engineering Conference (AUPEC). IEEE, 2014, hereinafter referenced as SINGH, in view of US 20160180220 A1, hereinafter referenced as BOETTCHER.
Regarding Claim 1
SINGH teaches:
An energy load prediction method, for energy supply users, based on a neural network, executed by a computing device, and comprising: (SINGH, p. 1, section I: These factors affect the nature of load profile, peak load variation in a particular area, e.g., in New South Wales (NSW), ... The proposed approach is evaluated using Mean of Mean Absolute Percentage Error (MMAPE), on load data for the NSW, Australia”;
SINGH, p. 2, section I: “The main focus of this paper is to investigate the performance of the proposed PRBFNN model for LF [load forecasting]. The results obtained show that, PRBFNN shows a considerable improvement in the forecasting accuracy, in comparison to existing ANN-based forecasting models, namely, FFNN, RBFNN, and ELMNN.”;
SINGH, p. 5, section V: “Therefore, to further increasing the forecasting accuracy, comparatively simpler but accurate forecasting model, i.e., RBFNN is optimized with PSO, and PRBFNN is proposed. Fig. 7 represents the visual comparison of the proposed PRBFNN models against the RBFNN and ELMNN models. No doubt, the forecasting profile obtained from PRBFNN seems very close to the actual profile.”;
Examiner’s Note: SINGH discloses the PRBFNN model (Particle Swarm Optimization-Based Radial Basis Function Neural Network) for predicting electric load for New South Wales, Australia (the residents of NSW correspond to recited “energy supply users”), where the implementation of a neural implies the use of a computing device to train and operate the neural network)
receiving a time period to be predicted; (SINGH, p. 5, section V: “Therefore, to further increasing the forecasting accuracy, comparatively simpler but accurate forecasting model, i.e., RBFNN is optimized with PSO, and PRBFNN is proposed. Fig. 7 represents the visual comparison of the proposed PRBFNN models against the RBFNN and ELMNN models. No doubt, the forecasting profile obtained from PRBFNN seems very close to the actual profile.”;
Examiner’s Note: SINGH, Fig. 7, shows that the predicted electric load was for a 72-hour time period from Jan. 4, 2011 at 00:00 to Jan. 6, 2011 at 24:00, corresponding to recited “time period to be predicted”)
and training and actually verifying a neural network model for predicting an energy load, wherein training and actually verifying a neural network model for predicting an energy load comprises training the neural network model based on a hybrid particle swarm optimization algorithm, (SINGH, p. 1, section I: “Therefore, a hybrid forecasting model, i.e., PSO-based RBFNN (PRBFNN), is proposed. The PSO in PRBFNN helps to obtain the optimal center width of Radial Basis Functions (RBFs), and weights of the output layer of RBFNN. The proposed approach is evaluated using Mean of Mean Absolute Percentage Error (MMAPE), on load data for the NSW, Australia [11].”;
SINGH, p. 2, section II.A: “In this paper, for the experimental study, real time hourly load data (in MWHrs.) of NSW State, Australia [12] is referred. To develop the load forecasting model, load data from 01 Jan’ 2010 to 06 Jan’ 2011 is considered. This data refers to 371 days, out of which, last three days data is reserved for testing the forecasting performance of ANN-based forecasting models. Initial 368 days data is utilized to build up the training and validation data set for ANN-based models.”;
Examiner’s Note: SINGH discloses the hybrid PRBFNN model (Particle Swarm Optimization-Based Radial Basis Function Neural Network) for predicting electric load for New South Wales, Australia, where the model was trained using data from a predetermined number of days (368 days) of actual data from New South Wales, Australia, and the model is evaluated (corresponding to recited “actually verified” using an error function based on real load data)
and actually verifying the neural network model based on n true (SINGH, p. 1, section I: “The proposed approach is evaluated using Mean of Mean Absolute Percentage Error (MMAPE), on load data for the NSW, Australia [11].”;
SINGH, p. 2, section II.A: “In this paper, for the experimental study, real time hourly load data (in MWHrs.) of NSW State, Australia [12] is referred. To develop the load forecasting model, load data from 01 Jan’ 2010 to 06 Jan’ 2011 is considered. This data refers to 371 days, out of which, last three days data is reserved for testing the forecasting performance of ANN-based forecasting models. Initial 368 days data is utilized to build up the training and validation data set for ANN-based models.”;
SINGH, p. 2, section III.A: “The optimal forecasting performance is measured and identified based on two error measures, i.e., Mean Absolute Percentage Error (MAPE)”;
Examiner’s Note: SINGH discloses data over 368 days in 2011 (corresponding to recited “n represents a load sampled number”) where the results are measured using MAPE which corresponds to “actually verifying the neural network model”)
without considering influences of weather, temperature and start-stop state on load (SINGH, p. 1, section 1: “A reduced input data set, only with electric load demand data and day-type parameter, is investigated, to observe the performances of various ANN models, without meteorological parameters.”;
SINGH, p. 2, section II.B: “For LF model inputs vector I(t) consists of 11 load values, i.e., 10 load values and one day-type parameter value”;
Examiner’s Note: SINGH discloses using a reduced input set with only electric load demand data and day-type parameters, and specifically excludes meteorological data (corresponding to recited “weather” and “temperature” exclusions), and there is no explicit disclosure of any data being considered that relates to a “start-stop state”)
inputting the time period into the neural network model for predicting an energy load, (SINGH, p. 1, section I: “A reduced input data set, only with electric load demand data and day-type parameter, is investigated, to observe the performances of various ANN models, without meteorological parameters.”;
SINGH, p. 2, section II.B: “For LF model inputs vector I(t) consists of 11 load values, i.e., 10 load values and one day-type parameter value”;
Examiner’s Note: SINGH explains that an input parameter to the NN is a day-type parameter (e.g., Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, or Sunday), because as explained in section II.A and depicted in Fig. 1, the day of the week impacts the electric load, with weekends having a lower electric load, and as shown in Fig. 7, a particular 3-day (or 72-hour) period is input into the neural network for prediction)
using the neural network model to output an energy load value in the time period so as to reduce a prediction deviation and improve an accuracy of energy load prediction (SINGH, p. 2, section 1: “The main focus of this paper is to investigate the performance of the proposed
PRBFNN model for LF. The results obtained show that, PRBFNN shows a considerable improvement in the forecasting accuracy”;
SINGH, p. 5, section V: “Therefore, to further increasing the forecasting accuracy, comparatively simpler but accurate forecasting model, i.e., RBFNN is optimized with PSO, and PRBFNN is proposed. Fig. 7 represents the visual comparison of the proposed PRBFNN models against the RBFNN and ELMNN models. No doubt, the forecasting profile obtained from PRBFNN seems very close to the actual profile.”;
Examiner’s Note: SINGH, Fig. 7, shows that the predicted electric load using the PRBFNN model was for a 72-hour time period from Jan. 4, 2011 at 00:00 to Jan. 6, 2011 at 24:00, corresponding to recited “time period”), where such PRBFNN model is shown to improve the accuracy of LF prediction (corresponding to recited “reduce a prediction deviation and improve an accuracy of energy load prediction” because improved accuracy means a reduced deviation between the predicted and true values)
However, SINGH fails to explicitly teach:
n true thermal load values and n thermal load predicted values ... n represents a thermal load sampled number
RMSE and DR
However, in a related field of endeavor (building energy consumption, see para. 0035), BOETTCHER teaches:
n true thermal load values and n thermal load predicted values ... n represents a thermal load sampled number (BOETTCHER, para. 0045: “Demand response layer 214 may be configured to optimize resource usage (e.g., electricity use, natural gas use, water use, etc.) and/or the monetary cost of such resource usage in response to satisfy the demand of building 100. ... Demand response layer 214 may receive inputs from other layers of BAS controller 202 (e.g., building subsystem integration layer 220, integrated control layer 218, etc.). The inputs received from other layers may include environmental or sensor inputs such as temperature, carbon dioxide levels, relative humidity levels, air quality sensor outputs, occupancy sensor outputs, room schedules, and the like. The inputs may also include inputs such as electrical use (e.g., expressed in kWh), thermal load measurements, pricing information, projected pricing, smoothed pricing, curtailment signals from utilities, and the like.”;
Examiner’s Note: BOETTCHER teaches optimizing electrical use with respect to thermal load measurements; the SINGH-BOETTCHER combination now uses the thermal load measurement data of BOETTCHER as additional training data for the PRBFNN model of SINGH)
through measurement indexes of experimental results which indexes include MAPE, RMSE and DR, (BOETTCHER, para. 0087: “ For example, model update module 328 may calculate a coefficient of variation of the root mean square error (CVRMSE) or any other metric that quantifies how well each set of model coefficients fits the equipment model.”;
BOETTCHER, para. 0092: “Outlier analysis module 410 may be configured to test data points and determine if a data point is reliable. For example, if a data point is more than a threshold (e.g., three standard deviations, four standard deviations, or another set value) away from the an expected value (e.g., the mean) of all of the data points, the data point may be determined as unreliable and discarded.”;
Examiner’s Note: BOETTCHER discloses data measurement metrics including RMSE and standard deviations (corresponding to recited “deviation rate”); the SINGH-BOETTCHER combination now uses the RMSE and standard deviation teachings of BOETTCHER to evaluate the results from the PRBFNN model of SINGH)
Before the effective filing date of the present application, it would have been obvious to combine the PRBFNN model of SINGH with the teachings of BOETTCHER as explained above. As disclosed by BOETTCHER, one of ordinary skill would have been motivated to do so in order to account for energy transfer rates “in and out of energy storage devices (e.g., thermal storage tanks, battery banks, etc.)”. (para. 0047). One of ordinary skill would further understand the benefit of using well-known data analysis metrics, including at least RMSE and standard deviations, because such metrics are well-known and accepted to those of ordinary skill in the art and in the scientific community.
Regarding Claim 7
Claim 7 claims a load prediction apparatus that corresponds to the method of claim 1, and is therefore rejected for the same reasons explained above with respect to claim 1.
Specifically, with respect to the “comprising a processor, wherein the processor is configured to implement each of the following modules located in the processor”, BOETTCHER discloses: “Processing circuit 204 is shown to include a processor 206 ... Memory 208 (e.g., memory, memory unit, storage device, etc.) may include one or more devices (e.g., RAM, ROM, Flash memory, hard disk storage, etc.) for storing data and/or computer code for completing or facilitating the various processes, layers and modules described in the present application.” (para. 0040)
Regarding Claim 9
BOETTCHER discloses:
A non-volatile memory, wherein a computer program is stored in the memory, and (BOETTCHER, para. 0040: “Memory 208 (e.g., memory, memory unit, storage device, etc.) may include one or more devices (e.g., RAM, ROM, Flash memory, hard disk storage, etc.) for storing data and/or computer code for completing or facilitating the various processes, layers and modules described in the present application. Memory 208 may be or include volatile memory or non-volatile memory.”)
The combination of SINGH and BOETTCHER discloses the following as explained with claim 1:
the computer program is configured to perform, when executed, implementing the load prediction method account to claim 1,
Regarding Claim 10
BOETTCHER discloses:
An electronic device, comprising a memory and a processor, wherein a computer program is stored in a non-volatile memory, (BOETTCHER, para. 0040: “Processing circuit 204 is shown to include a processor 206 and memory 208. Processor 206 can be implemented as a general purpose processor, an application specific integrated circuit (ASIC), one or more field programmable gate arrays (FPGAs), a group of processing components, or other suitable electronic processing components. Memory 208 (e.g., memory, memory unit, storage device, etc.) may include one or more devices (e.g., RAM, ROM, Flash memory, hard disk storage, etc.) for storing data and/or computer code for completing or facilitating the various processes, layers and modules described in the present application. Memory 208 may be or include volatile memory or non-volatile memory.”)
The combination of SINGH and BOETTCHER discloses the following as explained with claim 1:
and the processor is configured to execute the computer program to perform the load prediction method account to claim 1,
Claims 2-3, 8, 11-12, and 16-17 are rejected under 35 U.S.C. 103 as being unpatentable over SINGH in view of BOETTCHER and further in view of US 20190340462 A1, hereinafter referenced as PAO.
Regarding Claim 2
SINGH and BOETTCHER teach the method of claim 1 as explained above. SINGH further teaches:
wherein before the step of inputting the time period into the neural network model for predicting the energy load, the load prediction method further comprises: ... obtaining the neural network model by training using sample data. (SINGH, p. 2, section II: “Initial 368 days data is utilized to build up the training and validation data set for ANN-based models.”;
Examiner’s Note: SINGH teaches using sample data provided for NSW, Australia, to train the PRBFNN model, where such model is trained before it is used for inference)
However, SINGH and BOETTCHER fail to explicitly teach:
acquiring the neural network model from a third party
However, in a related field of endeavor (neural networks, see para. 0005), PAO explicitly teaches:
acquiring the neural network model from a third party (PAO, para. 0056: “While one or more embodiments described herein indicate that the object segmentation system 106 trains the convolutional neural network, in one or more embodiments, the object segmentation system 106 receives a trained convolutional neural network from a third-party device (e.g., a training server device).”;
Examiner’s Note: the SINGH-BOETTCHER- PAO combination now modifies the PRBFNN model of SINGH so that it is originally received from a third party server training device as in PAO).
Before the effective filing date of the present application, it would have been obvious to combine the PRBFNN model of SINGH with the teachings of BOETTCHER and PAO as explained above. One of ordinary skill would understand the benefit of utilizing a pre-trained model from a third party, because the heavy computations for training, and bug fixing and validation, would be done by the third party.
The examiner further notes that SINGH discloses receiving other types of models (e.g., classical models for comparison) that were developed by other parties. (see p. 4, section V).
Regarding Claim 3
SINGH, BOETTCHER, and PAO teach the method of claim 2 as explained above. SINGH further teaches:
determining a topology of an initial model, wherein the topology comprises: an input layer, a hidden layer, and an output layer; (SINGH, p. 3, section III.C: “As discussed earlier, RBFNN [20] necessarily contains only one hidden layer, and hence total three layers in whole. Therefore, this is considered as structure economic ANN. Input, hidden and output layer activation functions necessarily are, linear, radial basis and linear, respectively.”;
Examiner’s Note: The PRBFNN of SINGH is based on the RBFNN which has input, hidden, and output layers)
encoding parameters of the initial model to obtain an initial particle population, wherein the parameters comprise: (SINGH, p. 4 Table-I: Simulation Parameters for PRBFNN; Examiner’s Note: Table-I shows 320 initial particles in the swarm, where each parameter is given a parameter name in Table 1, and the weights of each node in the neural network similarly are given a value, corresponding to recited “encoding parameters of the initial model”)
a center parameter of a radial basis function, (SINGH, p. 1, section I: “The PSO in PRBFNN helps to obtain the optimal center width of Radial Basis Functions (RBFs), and weights of the output layer of RBFNN.”)
a variance parameter, (SINGH, p. 4, section IV: “Ic is the inertia that varies between 0.1 and 0.9.”)
a weight parameter of the hidden layer, and (SINGH, p. 3, Fig. 2; Examiner’s Note: Fig. 2 shows that the hidden layer in a FFNN has “hidden layer weights”)
a weight parameter of the output layer, and (SINGH, p. 1, section I: “The PSO in PRBFNN helps to obtain the optimal center width of Radial Basis Functions (RBFs), and weights of the output layer of RBFNN.”;
SINGH, p. 4, section IV: “This way, the widths of RBFs in hidden layer of RBFNN, and weights of the output layer are computed in PRBFNN model”)
each particle has at least one parameter; (SINGH, p. 4 Table-I: Simulation Parameters for PRBFNN; Examiner’s Note: Table-I shows 320 initial particles in the swarm, and parameters for the particles)
decoding the initial particle population to obtain initial parameters of the initial model; (SINGH, p. 4, section IV.A: “number of parameters to optimize is 16, problem dimension is fixed to 16. Swarm size is taken as 20 times of dimension, i.e., 320. Confidence parameters, c1 and c2, also termed as cognitive parameter and social parameter, respectively are provided a value of 2. Random numbers, r1 and r2 decide the diversity in the swarm or the solution group. Ic is the inertia that varies between 0.1 and 0.9. To make the final solution, very close to the global optimized value, number of iterations for the problem is fixed as 500, which also provide the stopping criteria for the optimization process.”; Examiner’s Note: SINGH discloses initializing the parameters to create the swarm of 320 particles given the constraints of Table-1)
assigning the initial parameters to the initial model to obtain an RBF network model; and (SINGH, p. 3, section III.C: “In RBFNN, weight matrices, Vrb and Wrb have weights of hidden layer, as well as output layer, respectively. As detailed in [12], the training of RBFNN primarily starts with the selection of centers of RBFs of hidden layer through unsupervised training or K-means clustering. Thereafter, the weights in the secondary layer are calculated. RBFNN structure, i.e., 11-88-1, is found the best LF model as discussed in [12].”; Examiner’s Note: SINGH discloses obtaining an initial RBFNN by initializing the weight matrices for the hidden and output layer)
optimizing the RBF network model by using a training sample and a test sample. (SINGH, p. 2, section II.A: “In this paper, for the experimental study, real time hourly load data (in MWHrs.) of NSW State, Australia [12] is referred. To develop the load forecasting model, load data from 01 Jan’ 2010 to 06 Jan’ 2011 is considered. This data refers to 371 days, out of which, last three days data is reserved for testing the forecasting performance of ANN-based forecasting models. Initial 368 days data is utilized to build up the training and validation data set for ANN-based models.”;
SINGH, p. 3, section III.C: “In RBFNN, weight matrices, Vrb and Wrb have weights of hidden layer, as well as output layer, respectively. As detailed in [12], the training of RBFNN primarily starts with the selection of centers of RBFs of hidden layer through unsupervised training or K-means clustering. Thereafter, the weights in the secondary layer are calculated. RBFNN structure, i.e., 11-88-1, is found the best LF model as discussed in [12].”;
Examiner’s Note: SINGH discloses obtaining an initial RBFNN by initializing the weight matrices for the hidden and output layer, and training such RBFNN using both test data and training data taken from real hourly load data from NSW, Australia)
Claim 8 depends from claim 7 and claims an apparatus that corresponds to the method of claim 3, and is therefore rejected for the same reasons explained above with respect to claims 3 and 7.
Claim 11 depends from claim 9 and claims a memory that corresponds to the method of claim 2, and is therefore rejected for the same reasons explained above with respect to claims 2 and 9.
Claim 12 depends from claim 11 and claims a memory that corresponds to the method of claim 3, and is therefore rejected for the same reasons explained above with respect to claims 3 and 11.
Claim 16 depends from claim 10 and claims an electronic device that corresponds to the method of claim 2, and is therefore rejected for the same reasons explained above with respect to claims 2 and 10.
Claim 17 depends from claim 16 and claims an electronic device that corresponds to the method of claim 3, and is therefore rejected for the same reasons explained above with respect to claims 3 and 16.
Claims 4-5, 13-14, and 18-19 are rejected under 35 U.S.C. 103 as being unpatentable over SINGH in view of BOETTCHER and PAO and further in view of US 20190242936 A1, hereinafter referenced as HE.
Regarding Claim 4
SINGH, BOETTCHER, and PAO teach the method of claim 3 as explained above. SINGH further teaches:
inputting the training sample and the test sample into the RBF network model respectively to obtain a test value and an expected value; (SINGH, p. 2, section II.A: “In this paper, for the experimental study, real time hourly load data (in MWHrs.) of NSW State, Australia [12] is referred. To develop the load forecasting model, load data from 01 Jan’ 2010 to 06 Jan’ 2011 is considered. This data refers to 371 days, out of which, last three days data is reserved for testing the forecasting performance of ANN-based forecasting models. Initial 368 days data is utilized to build up the training and validation data set for ANN-based models.”;
SINGH, p. 3, section III.C: “In RBFNN, weight matrices, Vrb and Wrb have weights of hidden layer, as well as output layer, respectively. As detailed in [12], the training of RBFNN primarily starts with the selection of centers of RBFs of hidden layer through unsupervised training or K-means clustering. Thereafter, the weights in the secondary layer are calculated. RBFNN structure, i.e., 11-88-1, is found the best LF model as discussed in [12].”;
Examiner’s Note: SINGH discloses training a RBFNN using both test data and training data taken from real hourly load data from NSW, Australia, where during the training stage, the training data is input and the output corresponds to the recited “test value”, and during testing and inference, the output corresponds to the recited “expected value”)
updating particles in the initial particle population by using the fitness value. (SINGH, p. 4, section IV.A: “A properly chosen fitness function plays a vital role in leading the whole optimization process towards global optimization. Eq. (13) denotes the fitness function as chosen in this work.”; Examiner’s Note: SINGH discloses using a fitness function as part of the particle swarm optimization process and using such fitness function when setting and refining the particle swarm)
However, SINGH, BOETTCHER, and PAO fail to explicitly teach:
selecting a norm of an error matrix consisting of the test value and the expected value as a fitness value
However, in a related field of endeavor (electrical circuit fault diagnoses, see para. 0002), HE teaches:
selecting a norm of an error matrix consisting of the test value and the expected value as a fitness value (HE, para. 0052: “for the fitness function, in the present invention, a norm of an error matrix between a prediction value and an expectation value of a prediction sample is selected as output of a target function”; Examiner’s Note: the SINGH-BOETTCHER- PAO-HE combination now modifies the fitness function of SINGH to utilize the norm of an error matrix as explained by HE)
Before the effective filing date of the present application, it would have been obvious to combine the PRBFNN model of SINGH with the teachings of BOETTCHER, PAO, and HE as explained above. As disclosed by HE, one of ordinary skill would have been motivated to do so “so that a residual between the prediction value and the expectation value is as small as possible when prediction is performed for the ... neural network.” (para. 0052).
Regarding Claim 5
SINGH, BOETTCHER, PAO, and HE teach the method of claim 4 as explained above. SINGH further teaches:
updating velocities and positions of the particles in the initial particle population; (SINGH, p. 4, section IV: “Each group (or swarm) member do have its own position as well as velocity in a direction, called local best position, posi(t) and local best velocity, vlocali(t)”)
updating an individual extremum of the particles in the initial particle population by using the fitness value, and (SINGH, p. 4, section IV: “Each group (or swarm) member do have its own position as well as velocity in a direction, called local best position, posi(t) and local best velocity, vlocali(t)”;
SINGH, p. 4, section IV.A: “A properly chosen fitness function plays a vital role in
leading the whole optimization process towards global optimization. Eq. (13) denotes the fitness function as chosen in this work. The basis of this choice is to minimize the forecasting error and increasing the forecasting performance of proposed model.”;
Examiner’s Note: local best velocity and local best position each correspond to recited “individual extremum of the particles”, and SINGH further teaches using the fitness function to optimize the particles, including the local best velocity/position)
updating a population extremum of the particles in the initial particle population by using the fitness value; and (SINGH, p. 4, section IV: “Best position amongst all local positions is treated as a global best position ... The movement of the swarm is decided by the global best position and global best velocity”;
SINGH, p. 4, section IV.A: “A properly chosen fitness function plays a vital role in
leading the whole optimization process towards global optimization. Eq. (13) denotes the fitness function as chosen in this work. The basis of this choice is to minimize the forecasting error and increasing the forecasting performance of proposed model.”;
Examiner’s Note: global best velocity and global best position each correspond to recited “population extremum of the particles”, and SINGH further teaches using the fitness function to optimize the particles, including the global best velocity/position)
mutating the particles in the initial particle population, and updating the particles when a fitness value of a new particle is better than a fitness value of an old particle. (SINGH, p. 4, section IV: “Based on the group velocity, the position of ith member can be modified by (12).
PNG
media_image4.png
24
406
media_image4.png
Greyscale
”;
SINGH, p. 4, section IV.A: “A properly chosen fitness function plays a vital role in
leading the whole optimization process towards global optimization. Eq. (13) denotes the fitness function as chosen in this work. The basis of this choice is to minimize the forecasting error and increasing the forecasting performance of proposed model.”;
Examiner’s Note: SINGH further teaches using the fitness function to optimize the particles, including the position of each particle, where each particle’s position is updated using the group velocity)
Claim 13 depends from claim 12 and claims a memory that corresponds to the method of claim 4, and is therefore rejected for the same reasons explained above with respect to claims 4 and 12.
Claim 14 depends from claim 13 and claims a memory that corresponds to the method of claim 5, and is therefore rejected for the same reasons explained above with respect to claims 5 and 13.
Claim 18 depends from claim 17 and claims an electronic device that corresponds to the method of claim 4, and is therefore rejected for the same reasons explained above with respect to claims 4 and 17.
Claim 19 depends from claim 18 and claims an electronic device that corresponds to the method of claim 5, and is therefore rejected for the same reasons explained above with respect to claims 5 and 18.
Claims 6, 15, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over SINGH, BOETTCHER, PAO, and HE and further in view of Han, Fei, et al. "An improved evolutionary extreme learning machine based on particle swarm optimization." Neurocomputing 116 (2013): 87-93, hereinafter referenced as HAN.
Regarding Claim 6
SINGH, BOETTCHER, PAO, and HE teach the method of claim 5 as explained above. However, SINGH, BOETTCHER, PAO, and HE fail to explicitly teach:
iteratively updating the velocities and the positions of the particles in the initial particle population by using the following formulas:
PNG
media_image3.png
348
688
media_image3.png
Greyscale
However, in a related field of endeavor (extreme learning machine based on particle swarm optimization), HAN teaches:
iteratively updating the velocities and the positions of the particles in the initial particle population by using the following formulas:
PNG
media_image3.png
348
688
media_image3.png
Greyscale
(HAN, p. 88, section 2.1: Then the original PSO [19,20] is described as
PNG
media_image5.png
140
410
media_image5.png
Greyscale
Examiner’s Note: the SINGH-BOETTCHER-PAO-HE-HAN combination now modifies the particle swarm optimization algorithms of SINGH to use the original particle swarm optimization formulations as shown in equations (4)-(5) of HAN)
Before the effective filing date of the present application, it would have been obvious to combine the PRBFNN model of SINGH with the teachings of BOETTCHER, PAO, HE, and HAN as explained above. As disclosed by HAN, one of ordinary skill would have been motivated to utilize the original particle swarm optimization functions because such functions have been adopted by others and have been peer-tested and peer-reviewed.
Claim 15 depends from claim 14 and claims a memory that corresponds to the method of claim 6, and is therefore rejected for the same reasons explained above with respect to claims 6 and 14.
Claim 20 depends from claim 19 and claims an electronic device that corresponds to the method of claim 6, and is therefore rejected for the same reasons explained above with respect to claims 6 and 19.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US 20110172897 A1 (Tsuzuki). “ According to the present embodiment, by using a neural network as the plant model, precise calculation having only a small modeling error can be performed with computational complexity that can be handled by an automobile computer. Although the neural network has many parameters, the parameters can be set at appropriate values by using an existing method, because the neural network does not identify these parameters as model parameters having physical meanings. (4) According to the present embodiment, the model adapter 9 adaptively corrects the combustion model so as to reduce the deviation of the predicted maximum cylinder pressure” (para. 0150).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MICHAEL C LEE whose telephone number is (571)272-4933. The examiner can normally be reached M-F 12:00 pm - 8:00 pm ET.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Omar Fernandez Rivas can be reached at 571-272-2589. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MICHAEL C. LEE/Examiner, Art Unit 2128
Prosecution Insights