CN117787444B - Intelligent algorithm rapid integration method and device for cluster countermeasure scene - Google Patents
Intelligent algorithm rapid integration method and device for cluster countermeasure scene Download PDFInfo
- Publication number
- CN117787444B CN117787444B CN202410211635.XA CN202410211635A CN117787444B CN 117787444 B CN117787444 B CN 117787444B CN 202410211635 A CN202410211635 A CN 202410211635A CN 117787444 B CN117787444 B CN 117787444B
- Authority
- CN
- China
- Prior art keywords
- algorithm
- integration
- target
- learner
- optimal position
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000010354 integration Effects 0.000 title claims abstract description 151
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000005457 optimization Methods 0.000 claims abstract description 41
- 238000004364 calculation method Methods 0.000 claims abstract description 26
- 230000007246 mechanism Effects 0.000 claims abstract description 21
- 238000012544 monitoring process Methods 0.000 claims abstract description 18
- 230000003993 interaction Effects 0.000 claims abstract description 17
- 230000008569 process Effects 0.000 claims abstract description 8
- 238000012549 training Methods 0.000 claims description 42
- 230000004927 fusion Effects 0.000 claims description 11
- 230000002159 abnormal effect Effects 0.000 claims description 9
- 230000001133 acceleration Effects 0.000 claims description 9
- 238000004891 communication Methods 0.000 claims description 8
- 238000013461 design Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 5
- 238000012360 testing method Methods 0.000 claims description 5
- 230000006870 function Effects 0.000 description 18
- 238000007726 management method Methods 0.000 description 17
- 230000000694 effects Effects 0.000 description 3
- 230000001965 increasing effect Effects 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013468 resource allocation Methods 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 229920001296 polysiloxane Polymers 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a cluster countermeasure scene-oriented intelligent algorithm rapid integration method and device, comprising the following steps: determining an algorithm integration target and designing an algorithm integration framework; executing corresponding integration steps according to the determined algorithm integration targets; the integration step comprises an integration step of the same type of algorithm, an integration step of different types of algorithms and an integration step of multi-objective algorithm optimization; designing an algorithm cooperation framework, implementing a cooperation mechanism, and continuously monitoring and adjusting interaction and cooperation among algorithms in the operation process of the cooperation mechanism; integrating a scheduling strategy and a management tool component into the same system, and deploying the scheduling strategy and the management tool component into an actual application environment to realize cooperative control of an algorithm; the running state, the calculation result and the resource occupation condition of the algorithm are monitored in real time, and real-time adjustment is carried out, so that the problems of system redundancy, low running efficiency, complex use and the like caused by the fact that the conventional algorithm integration method often lacks a unified management frame are solved.
Description
Technical Field
The application relates to the technical field of computers, in particular to an intelligent algorithm rapid integration method and device for a cluster countermeasure scene.
Background
With the widespread use of intelligent algorithms in many fields, more and more intelligent algorithms are used to solve various problems. To solve the complex problem, it is often necessary to integrate multiple algorithms together to form a unified system.
Since each algorithm relies on separate code implementation and management, there is a significant amount of redundant code in the system and the efficiency of operation is reduced, and furthermore, if there is no unified management mechanism, it can become complicated and time consuming to use and understand the algorithms. More seriously, if there is a lack of effective interaction and collaboration between algorithms, their overall performance may be impacted. The existing algorithm integration method often lacks a unified management framework, so that the problems of system redundancy, low operation efficiency, complex use and the like are caused.
Therefore, how to integrate algorithms effectively, improving the performance and efficiency of the system is an important technical challenge currently faced.
Disclosure of Invention
In the embodiment of the application, the intelligent algorithm rapid integration method for the cluster countermeasure scene solves the problems of system redundancy, low operation efficiency, complex use and the like caused by the fact that the conventional algorithm integration method often lacks a unified management frame.
In a first aspect, an embodiment of the present application provides a method for rapidly integrating an intelligent algorithm for a cluster countermeasure scenario, where the method includes: determining an algorithm integration target and designing an algorithm integration framework; wherein the algorithm integration framework is used for integrating different algorithms and supporting interaction and collaboration between the different algorithms; executing corresponding integration steps according to the determined algorithm integration targets; the integration step comprises an integration step of the same type of algorithm, an integration step of different types of algorithms and an integration step of multi-objective algorithm optimization; designing an algorithm cooperation framework, implementing a cooperation mechanism, and continuously monitoring and adjusting interaction and cooperation among algorithms in the operation process of the cooperation mechanism; integrating a scheduling strategy and a management tool component into the same system, and deploying the scheduling strategy and the management tool component into an actual application environment to realize cooperative control of an algorithm; and (3) monitoring the running state, the calculation result and the resource occupation condition of the algorithm in real time, and carrying out real-time adjustment.
With reference to the first aspect, in one possible implementation manner, the method further includes: when an error or abnormal condition is encountered in executing the corresponding integration step, an abnormal processing mechanism is used for processing the error or abnormal condition; after the integration step is completed, a test is performed to ensure that the communication between the algorithms is normal.
With reference to the first aspect, in a possible implementation manner, the step of integrating the algorithm of the same type includes: randomly sampling the data in the original data set used by the algorithm with a put back function to obtain a plurality of sub-data sets; wherein the accumulated result of the number of samples in each of the sub-data sets is the same as the number of samples in the original data set; training each sample in the sub-data set by using the same learner to obtain a first prediction result of each sample in the learner, and obtaining a trained first learner; when a new sample appears, inputting the new sample into the trained first learners, and obtaining a second prediction result of the new sample in each trained first learner; carrying out fusion operation of all the predicted results to obtain a final predicted result, thereby completing the integration step of the same type of algorithm; the calculation formula of the fusion operation is as follows: y=; Where y is the final predicted result,/>For the prediction result of the nth sample,/>And the weight corresponding to the prediction result of the nth sample.
With reference to the first aspect, in one possible implementation manner, the integrating step of the algorithm of the different types includes: executing an iteration step; wherein the iterative steps include: obtaining training samples of different types of algorithms in a data set of the training samples, and calculating a prediction error according to a prediction result of a current learner on the training samples and an actual value of the training samples so as to obtain a residual error; and adjusting the weight of the learner according to the obtained residual error to finish the iteration step, thereby finishing the integration step of algorithms of different types.
With reference to the first aspect, in one possible implementation manner, a calculation formula for obtaining the residual error is as follows: ; wherein/> Residual error obtained for iteration round c,/>For the prediction result of the current learner on training sample x,/>Representing the actual value of the training sample x.
With reference to the first aspect, in a possible implementation manner, the adjusting the weight of the learner according to the obtained residual includes: according to the formulaCalculating the weight obtained by the c-th iteration of the learner; wherein/>Weights obtained for learner round c iteration,/>Residual error obtained for iteration round c,/>Is a logic function, and represents that the value is 1 when the residual error is greater than 0, otherwise, the residual error is 0,/>Is a logic function, and represents that the value is 1 when the residual error is smaller than or equal to 0, otherwise, the residual error is 0; substituting the weight obtained by the learner in the c-th iteration and the prediction result of the current learner on the training sample x into the formula/>Obtaining a final prediction result, so as to adjust the weight of the learner according to the obtained residual error; wherein/>For the final prediction result,/>Weights obtained for learner round c iteration,/>The prediction result of the training sample x is the current learner.
With reference to the first aspect, in one possible implementation manner, the integrating step of the multi-objective algorithm optimization includes: a first updating step and a second updating step; wherein the first updating step includes: randomly generating a plurality of targets in the search space, wherein the position of each target corresponds to a multidimensional vector and has a multidimensional speed; wherein the search space is a set of potential solutions for the multi-objective algorithm; calculating the fitness value of each target; judging whether the fitness value of each target is better than the current individual optimal position and/or the global optimal position; if the fitness value of each target is better than the current individual optimal position and/or the global optimal position, updating the corresponding optimal position into the fitness value of each target; and if the fitness value of each target is not better than the current individual optimal position and/or the global optimal position, not updating the current individual optimal position and/or the global optimal position.
With reference to the first aspect, in a possible implementation manner, the second updating step includes: updating the speed and the position of the target; the calculation formula for updating the speed of the target is as follows: ; wherein/> For the speed of the target at the next moment, w is the inertial weight,/>Acceleration term weight learned for current individual optimal position,/>Acceleration term weights learned for globally optimal locations, rand () is a random function,/>For the speed of the target at the current moment, P is the current individual optimal position, G is the global optimal position,/>The position of the target at the current moment; the calculation formula for updating the position of the target is as follows: /(I); Wherein/>For the position of the target at the next moment,/>For the position of the target at the current moment,/>Which is the speed of the target at the next moment.
With reference to the first aspect, in a possible implementation manner, the integrating step of the multi-objective algorithm optimization further includes: judging whether the integration step of multi-objective algorithm optimization is finished according to the judgment rule; the judging rule comprises the following steps: judging whether a stopping condition is met; the stopping condition is that the preset maximum iteration times or the adaptability value is converged; if the stopping condition is not met, continuing to execute the integration step of multi-objective algorithm optimization; if the stopping condition is met, ending the integration step of the multi-objective algorithm optimization, and outputting the global optimal position, thereby completing the integration step of the multi-objective algorithm optimization.
In a second aspect, an embodiment of the present application provides an intelligent algorithm fast integration device for a cluster countermeasure scenario, where the device includes: the determining module is used for determining an algorithm integration target and designing an algorithm integration framework; wherein the algorithm integration framework is used for integrating different algorithms and supporting interaction and collaboration between the different algorithms; the execution module is used for executing corresponding integration steps according to the determined algorithm integration targets; the integration step comprises an integration step of the same type of algorithm, an integration step of different types of algorithms and an integration step of multi-objective algorithm optimization; the design module is used for designing an algorithm cooperation framework and implementing a cooperation mechanism, and continuously monitoring and adjusting interaction and cooperation among algorithms in the operation process of the cooperation mechanism; the integration module is used for integrating the scheduling strategy and the management tool component into the same system and deploying the scheduling strategy and the management tool component into an actual application environment to realize cooperative control of an algorithm; and the monitoring module is used for monitoring the running state of the algorithm, the calculation result and the resource occupation condition in real time and carrying out real-time adjustment.
With reference to the second aspect, in one possible implementation manner, the method further includes: when an error or abnormal condition is encountered in executing the corresponding integration step, an abnormal processing mechanism is used for processing the error or abnormal condition; after the integration step is completed, a test is performed to ensure that the communication between the algorithms is normal.
With reference to the second aspect, in a possible implementation manner, the step of integrating the algorithm of the same type includes: randomly sampling the data in the original data set used by the algorithm with a put back function to obtain a plurality of sub-data sets; wherein the accumulated result of the number of samples in each of the sub-data sets is the same as the number of samples in the original data set; training each sample in the sub-data set by using the same learner to obtain a first prediction result of each sample in the learner, and obtaining a trained first learner; when a new sample appears, inputting the new sample into the trained first learners, and obtaining a second prediction result of the new sample in each trained first learner; carrying out fusion operation of all the predicted results to obtain a final predicted result, thereby completing the integration step of the same type of algorithm; the calculation formula of the fusion operation is as follows: y=; Where y is the final predicted result,/>For the prediction result of the nth sample,/>And the weight corresponding to the prediction result of the nth sample.
With reference to the second aspect, in one possible implementation manner, the integrating step of the algorithm of the different types includes: executing an iteration step; wherein the iterative steps include: obtaining training samples of different types of algorithms in a data set of the training samples, and calculating a prediction error according to a prediction result of a current learner on the training samples and an actual value of the training samples so as to obtain a residual error; and adjusting the weight of the learner according to the obtained residual error to finish the iteration step, thereby finishing the integration step of algorithms of different types.
With reference to the second aspect, in one possible implementation manner, a calculation formula for obtaining the residual error is as follows: ; wherein/> Residual error obtained for iteration round c,/>For the prediction result of the current learner on training sample x,/>Representing the actual value of the training sample x.
With reference to the second aspect, in a possible implementation manner, the adjusting the weight of the learner according to the obtained residual includes: according to the formulaCalculating the weight obtained by the c-th iteration of the learner; wherein/>Weights obtained for learner round c iteration,/>Residual error obtained for iteration round c,/>Is a logic function, and represents that the value is 1 when the residual error is greater than 0, otherwise, the residual error is 0,/>Is a logic function, and represents that the value is 1 when the residual error is smaller than or equal to 0, otherwise, the residual error is 0; substituting the weight obtained by the learner in the c-th iteration and the prediction result of the current learner on the training sample x into the formula/>Obtaining a final prediction result, so as to adjust the weight of the learner according to the obtained residual error; wherein/>For the final prediction result,/>Weights obtained for learner round c iteration,/>The prediction result of the training sample x is the current learner.
With reference to the second aspect, in one possible implementation manner, the integrating step of the multi-objective algorithm optimization includes: a first updating step and a second updating step; wherein the first updating step includes: randomly generating a plurality of targets in the search space, wherein the position of each target corresponds to a multidimensional vector and has a multidimensional speed; wherein the search space is a set of potential solutions for the multi-objective algorithm; calculating the fitness value of each target; judging whether the fitness value of each target is better than the current individual optimal position and/or the global optimal position; if the fitness value of each target is better than the current individual optimal position and/or the global optimal position, updating the corresponding optimal position into the fitness value of each target; and if the fitness value of each target is not better than the current individual optimal position and/or the global optimal position, not updating the current individual optimal position and/or the global optimal position.
With reference to the second aspect, in a possible implementation manner, the second updating step includes: updating the speed and the position of the target; the calculation formula for updating the speed of the target is as follows: ; wherein/> For the speed of the target at the next moment, w is the inertial weight,/>Acceleration term weight learned for current individual optimal position,/>Acceleration term weights learned for globally optimal locations, rand () is a random function,/>For the speed of the target at the current moment, P is the current individual optimal position, G is the global optimal position,/>The position of the target at the current moment; the calculation formula for updating the position of the target is as follows: /(I); Wherein/>For the position of the target at the next moment,/>For the position of the target at the current moment,/>Which is the speed of the target at the next moment.
With reference to the second aspect, in one possible implementation manner, the integrating step of the multi-objective algorithm optimization further includes: judging whether the integration step of multi-objective algorithm optimization is finished according to the judgment rule; the judging rule comprises the following steps: judging whether a stopping condition is met; the stopping condition is that the preset maximum iteration times or the adaptability value is converged; if the stopping condition is not met, continuing to execute the integration step of multi-objective algorithm optimization; if the stopping condition is met, ending the integration step of the multi-objective algorithm optimization, and outputting the global optimal position, thereby completing the integration step of the multi-objective algorithm optimization.
One or more technical solutions provided in the embodiments of the present application at least have the following technical effects:
The embodiment of the application provides an intelligent algorithm rapid integration method for a cluster countermeasure scene, which designs an integration framework integrating different algorithms by determining an algorithm integration target, supports interaction and cooperation among the algorithms, provides a unified platform for algorithm integration, and enables the algorithm integration to be more efficient and convenient. According to the determined goal of algorithm integration, corresponding integration steps can be executed, so that the algorithm can be flexibly integrated and optimized according to actual requirements. The scheduling strategy and the management tool component are integrated into the same system and deployed into an actual application environment, so that the application under the cooperative control of the algorithm can be realized, the requirements of the actual scene can be better met, and the efficiency and the performance are improved. The running state, the calculation result and the resource occupation condition of the algorithm are monitored in real time, real-time adjustment and optimization can be carried out, potential problems can be found and solved in time, stable running and high-efficiency performance of the algorithm are ensured, and therefore the problems that the conventional algorithm integration method often lacks a unified management framework, and system redundancy, low running efficiency, complex use and the like are solved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the embodiments of the present invention or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a method for rapidly integrating intelligent algorithms for cluster countermeasure scenarios provided by an embodiment of the present application;
FIG. 2 is a flowchart showing the steps of integrating the same type of algorithm according to an embodiment of the present application;
FIG. 3 is a flowchart illustrating the steps for integrating different types of algorithms according to an embodiment of the present application;
FIG. 4 is a flowchart showing a first update procedure in the integration procedure of multi-objective algorithm optimization according to an embodiment of the present application;
FIG. 5 is a flowchart showing the steps involved in the integration of multi-objective algorithm optimization according to an embodiment of the present application;
Fig. 6 is a schematic diagram of an intelligent algorithm rapid integration device facing a cluster countermeasure scene provided by an embodiment of the present application;
Fig. 7 is a schematic diagram of an intelligent algorithm rapid integration server facing a cluster countermeasure scene according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. It will be apparent that the described embodiments are some, but not all, embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Some of the techniques involved in the embodiments of the present application are described below to aid understanding, and they should be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the application. Also, for the sake of clarity and conciseness, descriptions of well-known functions and constructions are omitted in the following description.
The embodiment of the application provides an intelligent algorithm rapid integration method for a cluster countermeasure scene, which comprises the steps of S101 to S107 as shown in fig. 1. In this embodiment, fig. 1 is only an execution sequence shown in the embodiment of the present application, and does not represent a unique execution sequence of a cluster-countermeasure-oriented intelligent algorithm rapid integration method, where the steps shown in fig. 1 may be executed in parallel or in reverse under a condition that a final result may be achieved.
S101: determining an algorithm integration target and designing an algorithm integration framework. Wherein the algorithm integration framework is used to integrate different algorithms and support interactions and collaboration between them.
Specifically, in algorithm integration, a target of explicit integration is first required. These goals may be tailored to the actual needs, such as improving prediction accuracy, reducing errors, enhancing generalization ability, etc. Once the goal is clear, an algorithm integration framework can be designed that aims to integrate the different algorithms and support interactions and collaboration between them. The framework should be able to accommodate various types of algorithms that are optimized and tuned to improve their performance in practical applications, and be able to be flexibly configured and tuned to achieve integration goals.
S102: and executing corresponding integration steps according to the determined algorithm integration targets. The integration step comprises an integration step of the same type of algorithm, an integration step of different types of algorithms and an integration step of multi-objective algorithm optimization.
In particular, overall performance may be improved by integrating multiple algorithms of the same type. For different types of algorithms, overall performance may be improved by integrating their respective advantages. For example, decision trees and neural networks may be integrated together, with the interpretability of the decision tree and the self-learning capabilities of the neural network being utilized to improve the performance of the model. For algorithms with multiple conflicting objectives, multi-objective algorithm optimization is required. For example, in a recommendation system, multiple objectives such as accuracy and diversity may need to be considered simultaneously. Through multi-objective optimization, a balance point can be found, so that each objective can be optimized. The integration step aims at integrating different algorithms, fully exploiting their advantages and achieving higher performance by appropriate interactions and collaboration. It should be noted that the implementation of each integration step should be performed based on the previously determined integration objective to ensure that the final integration result meets the actual requirements.
Fig. 2 is a specific flowchart of an integration step of the same type of algorithm according to an embodiment of the present application, as shown in fig. 2, including steps S201 to S204.
S201: and carrying out put-back random sampling on the data in the original data set used by the algorithm to obtain a plurality of sub-data sets. Wherein the accumulated result of the number of samples in each sub-data set is the same as the number of samples in the original data set.
In particular, the purpose of the above steps is to obtain a plurality of sub-data sets, the number of samples in each sub-data set being the same as the number of samples in the original data set, so that the diversity and representativeness of the data can be ensured while the problems of over-fitting and under-fitting are avoided.
S202: and training each sample in the sub-data set by using the same learner to obtain a first prediction result of each sample in the learner, and obtaining a trained first learner.
Specifically, the above steps can fully utilize the advantages of each learner, and improve the accuracy and stability of prediction.
S203: when a new sample appears, the new sample is input into the trained first learners, and a second prediction result of the new sample is obtained in each trained first learner.
S204: and carrying out fusion operation on all the predicted results to obtain a final predicted result, thereby completing the integration step of the same type of algorithm.
The calculation formula of the fusion operation is as follows: y=. Where y is the final predicted result,/>For the prediction result of the nth sample,/>And the weight corresponding to the prediction result of the nth sample.
Specifically, through the fusion operation, the prediction results of a plurality of learners can be comprehensively considered, and the accuracy and stability of prediction are further improved.
Fig. 3 is a specific flowchart of an integration step of different types of algorithms according to an embodiment of the present application, as shown in fig. 3, including steps S301 to S303.
S301: an iterative step is performed.
Specifically, before performing the iterative step, the number of iterations is typically set, during which the algorithm processes and learns the data multiple times to gradually improve its performance.
S302: and acquiring training samples of different types of algorithms in the data set, and calculating a prediction error according to a prediction result of the current learner on the training samples and an actual value of the training samples, so as to acquire a residual error.
Further, the calculation formula for obtaining the residual is as follows: . Wherein/> Residual error obtained for iteration round c,/>For the prediction result of the current learner on training sample x,/>Representing the actual value of the training sample x.
S303: and adjusting the weight of the learner according to the obtained residual error to finish the iteration step, thereby finishing the integration step of algorithms of different types.
Further, adjusting the weight of the learner according to the obtained residual error, including: according to the formulaThe weights obtained by the learner in the c-th iteration are calculated. Wherein/>Weights obtained for learner round c iteration,/>Residual error obtained for iteration round c,/>Is a logic function, and represents that the value is 1 when the residual error is greater than 0, otherwise, the residual error is 0,/>Is a logic function, and represents that the value is 1 when the residual error is smaller than or equal to 0, and is 0 otherwise.
Specifically, if the residual is greater than 0, the weight is reduced, helping to reduce the complexity of the model and avoiding over-fitting. Because a large residual means that the prediction results of the model differ significantly from the actual values, this may be because the model is too complex, overfitting the training data, resulting in poor performance on new, unseen data. By reducing the weights, the model can be simplified, reducing the risk of overfitting. If the residual is less than or equal to 0, the weight will increase, helping to improve the prediction accuracy of the model. When the residual error is smaller, the predicted result of the model is closer to the actual value, and the weight can be increased to strengthen the good performance, so that the model can make more accurate prediction under similar situations. The strategy for dynamically adjusting the weights is beneficial to gradually optimizing the model in the training process, and improves generalization capability and prediction accuracy of the model. The model can be better adapted to various data distribution, and the prediction capability of the model on unknown data is enhanced. At the same time, this also makes the model more robust, reducing the risk of overfitting. The model in the present application refers to a learner.
Further, the weight obtained by the c-th iteration of the learner and the predicted result of the current learner on the training sample x are substituted into the formulaAnd obtaining a final prediction result, so that the weight of the learner is adjusted according to the obtained residual error. Wherein/>For the final prediction result,/>Weights obtained for learner round c iteration,/>The prediction result of the training sample x is the current learner.
Specifically, in each iteration, a residual is calculated from the current prediction result and the actual value, and then the weight of the model is adjusted based on this residual. In this way, the algorithm can gradually learn and optimize to improve its prediction accuracy.
Specifically, the integration step of the multi-objective algorithm optimization includes: a first updating step and a second updating step.
Fig. 4 is a specific flowchart of a first updating step in the integration step of the multi-objective algorithm optimization according to the embodiment of the present application, as shown in fig. 4, including steps S401 to S405.
S401: a plurality of objects are randomly generated in the search space, each object having a position corresponding to a multidimensional vector and having a multidimensional velocity. Where the search space is a collection of potential solutions for the multi-objective algorithm.
S402: and calculating the fitness value of each target. In particular, the fitness value represents the performance of the objective in the optimization problem.
S403: and judging whether the fitness value of each target is better than the current individual optimal position and/or the global optimal position.
If the fitness value of each target is better than the current individual optimal position and/or the global optimal position, that is, if the judgment result is yes, step S404 is executed: and updating the corresponding optimal position to the fitness value of each target.
If the fitness value of each target is not better than the current individual optimal position and/or the global optimal position, that is, the judgment result is no, step S405 is executed: no update of the current individual optimal position and/or global optimal position is performed.
Specifically, the second updating step in the integration step of the multi-objective algorithm optimization includes: the speed and position of the target are updated.
The calculation formula for updating the speed of the target is as follows: . Wherein/> For the speed of the target at the next moment, w is the inertial weight,/>Acceleration term weight learned for current individual optimal position,/>Acceleration term weights learned for globally optimal locations, rand () is a random function,/>For the speed of the target at the current moment, P is the current individual optimal position, G is the global optimal position,/>Is the position of the target at the current moment.
Specifically, the inertia weight w determines the balance of the target between the original speed and the changed speed, a larger value of the inertia weight w results in the target being more prone to maintain the original speed, and a smaller value of the inertia weight w results in the target being more prone to change its speed. The inertial weight w may decrease linearly with increasing number of iterations, and may also be adaptively adjusted according to the current state of the algorithm or other parameters. For example, when the algorithm falls into a local optimum, the value of the inertia weight w may be reduced to enhance the effect of the learning term, helping the algorithm jump out of the local optimum. The value of the inertial weight w may also be determined randomly, which may help to enhance the exploratory capabilities of the algorithm, but requires careful control of the random range to avoid adversely affecting the performance of the algorithm. rand () is a random function that generates a random number between 0 and 1 that increases the randomness of the target, making it possible to explore new, untried regions of the search space. The object of the present application is to correspond to particles. Acceleration term weight for learning optimal position of current individualAcceleration term weight/>, for global optimal position learningThe importance of the individual optimal position and the global optimal position is determined. Greater/>The value means approaching the individual's optimal location faster, while a larger/>The value then means approaching the global optimum faster. By adjusting the two weights, the utilization degree of the algorithm on the individual and global information can be balanced so as to adapt to different optimization problems. Generally, as the second updating step proceeds, it is desirable that the target pay more attention to the global optimum position, i.e., the global optimum information, thereby gradually reducing the dependence on the individual optimum position, i.e., the individual information. Thus, the/>, is adjusted or reduced during the second updating stepWhile increasing or maintaining a higher/>A value that generally gives a better update effect.
The calculation formula for updating the position of the target is as follows: . Wherein, For the position of the target at the next moment,/>For the position of the target at the current moment,/>Which is the speed of the target at the next moment. Specifically, a calculation formula for updating the position of the target adds the speed of the target at the next time to the position of the target at the current time, thereby obtaining the position of the target at the next time.
Further, by effectively updating the speed and position of the target, and using preset stopping conditions, a satisfactory solution can be found in a reasonable time.
Fig. 5 is a specific flowchart of steps further included in the integration step of the multi-objective algorithm optimization according to the embodiment of the present application, as shown in fig. 5, including steps S501 to S504.
S501: and judging whether the integration step of the multi-objective algorithm optimization is finished according to the judgment rule.
S502: whether a stop condition is satisfied is determined. The stopping condition is that the preset maximum iteration times or the adaptability value is converged.
If the stop condition is not satisfied, that is, if the determination result is no, step S503 is executed: and continuing to execute the integration step of the multi-objective algorithm optimization.
If the stop condition is satisfied, that is, if the determination result is yes, step S504 is executed: and finishing the integration step of the multi-objective algorithm optimization, and outputting the global optimal position, thereby finishing the integration step of the multi-objective algorithm optimization.
S103: and designing an algorithm cooperation framework, implementing a cooperation mechanism, and continuously monitoring and adjusting interaction and cooperation among algorithms in the operation process of the cooperation mechanism.
Specifically, the steps enable different algorithms to effectively cooperate by designing an algorithm cooperation framework. And in particular may relate to defining communication protocols, determining data sharing manners, formulating work allocation policies, and the like. The advantages of performing the above steps are: 1. and the efficiency is improved. By reasonably dividing work, each algorithm can work in parallel, and the overall efficiency is improved. 2. And (5) optimizing resources. By reasonable resource allocation, system resources can be maximally utilized. 3. Robustness is enhanced. The cooperation of multiple algorithms can enhance the robustness of the system and cope with faults or anomalies that may occur in a single algorithm.
S104: the scheduling strategy and the management tool component are integrated into the same system and deployed into an actual application environment, so that cooperative control of the algorithm is realized.
In particular, the above steps integrate tools and policies for managing algorithm execution and scheduling into the same system for centralized management and monitoring. The advantages of performing the above steps are: 1. simplifying management. The centralized management tool can simplify the operation and reduce the management complexity. 2. The efficiency is improved. By centralized scheduling, tasks and resources can be more efficiently allocated, thereby improving overall efficiency. 3. And the response speed is improved. And various tasks and events are responded quickly, and the system is ensured to run stably.
S105: and (3) monitoring the running state, the calculation result and the resource occupation condition of the algorithm in real time, and carrying out real-time adjustment.
Specifically, the steps can find problems in time and adjust. The advantages of performing the above steps are: 1. and timely finding out abnormality. Real-time monitoring is helpful for timely finding out abnormality or error and avoiding problem expansion. 2. Optimizing performance: according to the monitoring data, the algorithm can be optimized, and the performance of the algorithm is improved. 3. And (3) reasonably distributing resources: according to the occupation condition of the resources, the resource allocation can be reasonably adjusted, and the resource waste or shortage is avoided.
S106: when an error or exception condition is encountered while the corresponding integration step is performed, the error or exception condition is handled using an exception handling mechanism.
Specifically, the exception handling mechanism in the above steps may be a preset exception handling mechanism. The advantages of performing the above steps are: 1. system stability is enhanced. Through effective exception handling, the risk of a system crashing due to errors or exceptions can be reduced. 2. And (5) quick recovery. The rapid identification and handling of anomalies can help the system quickly resume normal operating conditions. 3. The loss is reduced. Handling exceptions in time may reduce potential losses due to system failures.
S107: after the integration step is completed, a test is performed to ensure that the communication between the algorithms is normal.
Specifically, the above steps are performed after the integration step is completed. The advantages of performing the above steps are: 1. ensuring stable communication. Through testing, stable and reliable communication between algorithms can be ensured. 2. Problems were found in advance. Finding problems and repairing them before formal deployment can reduce the later maintenance costs. 3. And the user satisfaction degree is improved. A stable and efficient integrated system may provide a better use experience for the user.
The embodiment of the application also provides an intelligent algorithm rapid integration device 600 for cluster countermeasure scene, as shown in fig. 6, the device comprises: a determination module 601, an execution module 602, a design module 603, an integration module 604, and a monitoring module 605.
The determining module 601 is configured to determine an objective of algorithm integration and design an algorithm integration framework. Wherein the algorithm integration framework is used to integrate different algorithms and support interactions and collaboration between them.
An execution module 602, configured to execute a corresponding integration step according to the determined target of algorithm integration. The integration step comprises an integration step of the same type of algorithm, an integration step of different types of algorithms and an integration step of multi-objective algorithm optimization.
The design module 603 is configured to design an algorithm collaboration framework and implement a collaboration mechanism, and continuously monitor and adjust interactions and collaboration between algorithms during operation of the collaboration mechanism.
And the integrating module 604 is used for integrating the scheduling strategy and the management tool component into the same system and deploying the scheduling strategy and the management tool component into an actual application environment to realize cooperative control of the algorithm.
The monitoring module 605 is configured to monitor the running state, the calculation result and the resource occupation condition of the algorithm in real time, and perform real-time adjustment.
Some of the modules of the apparatus of the present application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, classes, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
The apparatus or module set forth in the embodiments of the application may be implemented in particular by a computer chip or entity, or by a product having a certain function. For convenience of description, the above devices are described as being functionally divided into various modules, respectively. The functions of each module may be implemented in the same piece or pieces of software and/or hardware when implementing the embodiments of the present application. Of course, a module that implements a certain function may be implemented by a plurality of sub-modules or a combination of sub-units.
The methods, apparatus or modules described in this application may be implemented in computer readable program code means and in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer readable medium storing computer readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, application specific integrated circuits (english: application SPECIFIC INTEGRATED Circuit; ASIC), programmable logic controller and embedded microcontroller, examples of the controller including but not limited to the following microcontrollers: ARC 625D, atmel AT91SAM, microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic of the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller in a pure computer readable program code, it is well possible to implement the same functionality by logically programming the method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers, etc. Such a controller can be regarded as a hardware component, and means for implementing various functions included therein can also be regarded as a structure within the hardware component. Or even means for achieving the various functions may be regarded as either software modules implementing the methods or structures within hardware components.
As shown in fig. 7, the embodiment of the present application further provides an intelligent algorithm rapid integration server for a cluster countermeasure scene, which includes a memory 701 and a processor 702; memory 701 is used to store computer-executable instructions; the processor 702 is configured to execute computer executable instructions to implement the intelligent algorithm rapid integration method for cluster countermeasure scenarios according to the embodiment of the present application.
The embodiment of the application also provides a computer readable storage medium which stores executable instructions, and the computer can realize the intelligent algorithm rapid integration method facing the cluster countermeasure scene.
From the above description of embodiments, it will be apparent to those skilled in the art that the present application may be implemented in software plus necessary hardware. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product or may be embodied in the implementation of data migration. The computer software product may be stored on a storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and include instructions for causing a computer device (which may be a personal computer, mobile terminal, server, or network device, etc.) to perform the methods described in the embodiments of the present application.
In this specification, each embodiment is described in a progressive manner, and the same or similar parts of each embodiment are referred to each other, and each embodiment is mainly described as a difference from other embodiments. All or portions of the present application are operational with numerous general purpose or special purpose computer system environments or configurations.
The above embodiments are only for illustrating the technical solution of the present application, and not for limiting the present application; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced with equivalents; such modifications and substitutions do not depart from the spirit of the application.
Claims (7)
1. The intelligent algorithm rapid integration method for the cluster countermeasure scene is characterized by comprising the following steps of:
Determining an algorithm integration target and designing an algorithm integration framework; wherein the algorithm integration framework is used for integrating different algorithms and supporting interaction and collaboration between the different algorithms;
Executing corresponding integration steps according to the determined algorithm integration targets; the integration step comprises an integration step of the same type of algorithm, an integration step of different types of algorithms and an integration step of multi-objective algorithm optimization;
the step of integrating the same type of algorithm comprises:
Randomly sampling the data in the original data set used by the algorithm with a put back function to obtain a plurality of sub-data sets; wherein the accumulated result of the number of samples in each of the sub-data sets is the same as the number of samples in the original data set;
Training each sample in the sub-data set by using the same learner to obtain a first prediction result of each sample in the learner, and obtaining a trained first learner;
When a new sample appears, inputting the new sample into the trained first learners, and obtaining a second prediction result of the new sample in each trained first learner;
Carrying out fusion operation of all the predicted results to obtain a final predicted result, thereby completing the integration step of the same type of algorithm;
the calculation formula of the fusion operation is as follows:
; where y is the final predicted result,/> For the prediction result of the nth sample,/>The weight corresponding to the prediction result of the nth sample;
The step of integrating the different types of algorithms comprises the following steps: executing an iteration step;
Wherein the iterative steps include: obtaining training samples of different types of algorithms in a data set of the training samples, and calculating a prediction error according to a prediction result of a current learner on the training samples and an actual value of the training samples so as to obtain a residual error;
According to the obtained residual error, the weight of the learner is adjusted to complete the iteration step, so that the integration step of algorithms of different types is completed;
The integration step of the multi-objective algorithm optimization comprises the following steps:
A first updating step and a second updating step;
wherein the first updating step includes:
Randomly generating a plurality of targets in the search space, wherein the position of each target corresponds to a multidimensional vector and has a multidimensional speed; wherein the search space is a set of potential solutions for the multi-objective algorithm;
calculating the fitness value of each target;
judging whether the fitness value of each target is better than the current individual optimal position and/or the global optimal position;
If the fitness value of each target is better than the current individual optimal position and/or the global optimal position, updating the corresponding optimal position into the fitness value of each target;
If the fitness value of each target is not better than the current individual optimal position and/or the global optimal position, the current individual optimal position and/or the global optimal position are not updated;
Designing an algorithm cooperation framework, implementing a cooperation mechanism, and continuously monitoring and adjusting interaction and cooperation among algorithms in the operation process of the cooperation mechanism;
integrating a scheduling strategy and a management tool component into the same system, and deploying the scheduling strategy and the management tool component into an actual application environment to realize cooperative control of an algorithm;
And (3) monitoring the running state, the calculation result and the resource occupation condition of the algorithm in real time, and carrying out real-time adjustment.
2. The intelligent algorithm rapid integration method for cluster countermeasure scene according to claim 1, further comprising:
When an error or abnormal condition is encountered in executing the corresponding integration step, an abnormal processing mechanism is used for processing the error or abnormal condition;
after the integration step is completed, a test is performed to ensure that the communication between the algorithms is normal.
3. The intelligent algorithm rapid integration method for cluster countermeasure scenes according to claim 1, wherein the calculation formula for obtaining the residual is as follows:
; wherein/> Residual error obtained for iteration round c,/>For the prediction result of the current learner on training sample x,/>Representing the actual value of the training sample x.
4. The intelligent algorithm rapid integration method for cluster countermeasure scene according to claim 3, wherein the adjusting the weight of the learner according to the obtained residual error comprises:
According to the formula Calculating the weight obtained by the c-th iteration of the learner; wherein/>Weights obtained for learner round c iteration,/>Residual error obtained for iteration round c,/>Is a logic function, and represents that the value is 1 when the residual error is greater than 0, otherwise, the residual error is 0,/>Is a logic function, and represents that the value is 1 when the residual error is smaller than or equal to 0, otherwise, the residual error is 0;
Substituting the weight obtained by the learner in the c-th iteration and the prediction result of the current learner on the training sample x into a formula Obtaining a final prediction result, so as to adjust the weight of the learner according to the obtained residual error; wherein/>For the final prediction result,/>Weights obtained for learner round c iteration,/>The prediction result of the training sample x is the current learner.
5. The intelligent algorithm rapid integration method for cluster countermeasure scene according to claim 1, wherein the second updating step includes:
Updating the speed and the position of the target;
the calculation formula for updating the speed of the target is as follows:
; wherein, For the speed of the target at the next moment, w is the inertial weight,/>Acceleration term weight learned for current individual optimal position,/>Acceleration term weights learned for globally optimal locations, rand () is a random function,/>For the speed of the target at the current moment, P is the current individual optimal position, G is the global optimal position,/>The position of the target at the current moment;
the calculation formula for updating the position of the target is as follows:
; wherein/> For the position of the target at the next moment,/>For the position of the target at the current moment,/>Which is the speed of the target at the next moment.
6. The intelligent algorithm rapid integration method for cluster countermeasure scene according to claim 1, wherein the integration step of multi-objective algorithm optimization further comprises:
Judging whether the integration step of multi-objective algorithm optimization is finished according to the judgment rule;
The judging rule comprises the following steps:
Judging whether a stopping condition is met; the stopping condition is that the preset maximum iteration times or the adaptability value is converged;
if the stopping condition is not met, continuing to execute the integration step of multi-objective algorithm optimization;
If the stopping condition is met, ending the integration step of the multi-objective algorithm optimization, and outputting the global optimal position, thereby completing the integration step of the multi-objective algorithm optimization.
7. An intelligent algorithm rapid integration device for a cluster countermeasure scene is characterized by comprising:
The determining module is used for determining an algorithm integration target and designing an algorithm integration framework; wherein the algorithm integration framework is used for integrating different algorithms and supporting interaction and collaboration between the different algorithms;
the execution module is used for executing corresponding integration steps according to the determined algorithm integration targets; the integration step comprises an integration step of the same type of algorithm, an integration step of different types of algorithms and an integration step of multi-objective algorithm optimization; the step of integrating the same type of algorithm comprises: randomly sampling the data in the original data set used by the algorithm with a put back function to obtain a plurality of sub-data sets; wherein the accumulated result of the number of samples in each of the sub-data sets is the same as the number of samples in the original data set; training each sample in the sub-data set by using the same learner to obtain a first prediction result of each sample in the learner, and obtaining a trained first learner; when a new sample appears, inputting the new sample into the trained first learners, and obtaining a second prediction result of the new sample in each trained first learner; carrying out fusion operation of all the predicted results to obtain a final predicted result, thereby completing the integration step of the same type of algorithm; the calculation formula of the fusion operation is as follows: ; where y is the final predicted result,/> For the prediction result of the nth sample,/>The weight corresponding to the prediction result of the nth sample; the step of integrating the different types of algorithms comprises the following steps: executing an iteration step; wherein the iterative steps include: obtaining training samples of different types of algorithms in a data set of the training samples, and calculating a prediction error according to a prediction result of a current learner on the training samples and an actual value of the training samples so as to obtain a residual error; according to the obtained residual error, the weight of the learner is adjusted to complete the iteration step, so that the integration step of algorithms of different types is completed; the integration step of the multi-objective algorithm optimization comprises the following steps: a first updating step and a second updating step; wherein the first updating step includes: randomly generating a plurality of targets in the search space, wherein the position of each target corresponds to a multidimensional vector and has a multidimensional speed; wherein the search space is a set of potential solutions for the multi-objective algorithm; calculating the fitness value of each target; judging whether the fitness value of each target is better than the current individual optimal position and/or the global optimal position; if the fitness value of each target is better than the current individual optimal position and/or the global optimal position, updating the corresponding optimal position into the fitness value of each target; if the fitness value of each target is not better than the current individual optimal position and/or the global optimal position, the current individual optimal position and/or the global optimal position are not updated;
the design module is used for designing an algorithm cooperation framework and implementing a cooperation mechanism, and continuously monitoring and adjusting interaction and cooperation among algorithms in the operation process of the cooperation mechanism;
the integration module is used for integrating the scheduling strategy and the management tool component into the same system and deploying the scheduling strategy and the management tool component into an actual application environment to realize cooperative control of an algorithm;
And the monitoring module is used for monitoring the running state of the algorithm, the calculation result and the resource occupation condition in real time and carrying out real-time adjustment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410211635.XA CN117787444B (en) | 2024-02-27 | 2024-02-27 | Intelligent algorithm rapid integration method and device for cluster countermeasure scene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410211635.XA CN117787444B (en) | 2024-02-27 | 2024-02-27 | Intelligent algorithm rapid integration method and device for cluster countermeasure scene |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117787444A CN117787444A (en) | 2024-03-29 |
| CN117787444B true CN117787444B (en) | 2024-05-17 |
Family
ID=90396694
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410211635.XA Active CN117787444B (en) | 2024-02-27 | 2024-02-27 | Intelligent algorithm rapid integration method and device for cluster countermeasure scene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117787444B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113326882A (en) * | 2021-05-31 | 2021-08-31 | 浪潮通用软件有限公司 | Model integration method and device based on classification and regression algorithm |
| CN114139696A (en) * | 2020-09-03 | 2022-03-04 | 顺丰科技有限公司 | Model processing method and device based on algorithm integration platform and computer equipment |
| WO2023087953A1 (en) * | 2021-11-22 | 2023-05-25 | 华为技术有限公司 | Method and apparatus for searching for neural network ensemble model, and electronic device |
| CN116167415A (en) * | 2023-02-27 | 2023-05-26 | 清华大学深圳国际研究生院 | Policy decision method in multi-agent cooperation and antagonism |
| CN116340839A (en) * | 2023-02-08 | 2023-06-27 | 北京大数据先进技术研究院 | Algorithm selecting method and device based on ant lion algorithm |
| CN116955959A (en) * | 2023-08-11 | 2023-10-27 | 中科(厦门)数据智能研究院 | Time sequence prediction integration method based on multi-objective evolution algorithm |
-
2024
- 2024-02-27 CN CN202410211635.XA patent/CN117787444B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114139696A (en) * | 2020-09-03 | 2022-03-04 | 顺丰科技有限公司 | Model processing method and device based on algorithm integration platform and computer equipment |
| CN113326882A (en) * | 2021-05-31 | 2021-08-31 | 浪潮通用软件有限公司 | Model integration method and device based on classification and regression algorithm |
| WO2023087953A1 (en) * | 2021-11-22 | 2023-05-25 | 华为技术有限公司 | Method and apparatus for searching for neural network ensemble model, and electronic device |
| CN116340839A (en) * | 2023-02-08 | 2023-06-27 | 北京大数据先进技术研究院 | Algorithm selecting method and device based on ant lion algorithm |
| CN116167415A (en) * | 2023-02-27 | 2023-05-26 | 清华大学深圳国际研究生院 | Policy decision method in multi-agent cooperation and antagonism |
| CN116955959A (en) * | 2023-08-11 | 2023-10-27 | 中科(厦门)数据智能研究院 | Time sequence prediction integration method based on multi-objective evolution algorithm |
Non-Patent Citations (2)
| Title |
|---|
| 一种求解集成生产计划的混合协同进化算法;周泓;王建;谭小卫;;计算机集成制造系统;20070715(07);全文 * |
| 集成化服务链多目标全局优化模型与算法;吴映波;王旭;刘昕;;重庆大学学报;20120815(08);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117787444A (en) | 2024-03-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107231436B (en) | Method and device for scheduling service | |
| Etemadi et al. | A cost-efficient auto-scaling mechanism for IoT applications in fog computing environment: a deep learning-based approach | |
| CN113760553B (en) | Mixed part cluster task scheduling method based on Monte Carlo tree search | |
| Khan et al. | Optimizing hadoop parameter settings with gene expression programming guided PSO | |
| CN116680062B (en) | Application scheduling deployment method based on big data cluster and storage medium | |
| CN113467944B (en) | Resource deployment device and method for complex software system | |
| CN112422699A (en) | Unmanned aerial vehicle cluster action scheme generation method based on dynamic adjustment | |
| Naik et al. | Performance improvement of MapReduce framework in heterogeneous context using reinforcement learning | |
| CN116047934A (en) | Real-time simulation method and system for unmanned aerial vehicle cluster and electronic equipment | |
| CN115552412A (en) | Graph convolution reinforcement learning by utilizing heterogeneous agent group | |
| Gand et al. | A fuzzy controller for self-adaptive lightweight edge container orchestration | |
| Wu et al. | Traffic signal networks control optimize with PSO algorithm | |
| WO2023089350A1 (en) | An architecture for a self-adaptive computation management in edge cloud | |
| CN117787444B (en) | Intelligent algorithm rapid integration method and device for cluster countermeasure scene | |
| CN116204849A (en) | Data and model fusion method for digital twin application | |
| CN106874215B (en) | Serialized storage optimization method based on Spark operator | |
| CN115941696A (en) | Heterogeneous Big Data Distributed Cluster Storage Optimization Method | |
| CN115759979A (en) | Process intelligent processing method and system based on RPA and process mining | |
| Skarin et al. | An assisting model predictive controller approach to control over the cloud | |
| Wang et al. | An estimation of distribution algorithm for the flexible job-shop scheduling problem | |
| CN114492052A (en) | Global stream level network simulation method, system and device | |
| CN112926952A (en) | Cloud computing-combined big data office business processing method and big data server | |
| CN112925831A (en) | Big data mining method and big data mining service system based on cloud computing service | |
| CN108009686B (en) | Photovoltaic power generation power prediction method, device and system | |
| Singh et al. | A GA based job scheduling strategy for computational grid |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |