CN112306859A - Improved software self-adaptive testing method - Google Patents
Improved software self-adaptive testing method Download PDFInfo
- Publication number
- CN112306859A CN112306859A CN202010663807.9A CN202010663807A CN112306859A CN 112306859 A CN112306859 A CN 112306859A CN 202010663807 A CN202010663807 A CN 202010663807A CN 112306859 A CN112306859 A CN 112306859A
- Authority
- CN
- China
- Prior art keywords
- test
- theta
- input
- software
- decision
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3668—Software testing
- G06F11/3672—Test management
- G06F11/3684—Test management for test design, e.g. generating new test cases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3668—Software testing
- G06F11/3672—Test management
- G06F11/3688—Test management for test execution, e.g. scheduling of test suites
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Debugging And Monitoring (AREA)
Abstract
The invention relates to an improved software self-adaptive testing method, which is based on a software control theory model and aims at improving the traditional self-adaptive random testing method. According to key problems faced by software testing, the invention can effectively solve the problem of software input space combination explosion by improving the hypothesis of the traditional software self-adaptive testing method, fully utilizes intermediate testing data generated in the testing process, realizes dynamic and self-adaptive feedback regulation of the software testing process, can effectively reduce the number of test cases, improves the testing efficiency and reduces the testing cost. Therefore, the method has important engineering application value.
Description
Technical Field
The invention belongs to the technical field of software engineering and software testing, and particularly relates to an improved software self-adaptive testing method.
Background
With the rapid development of information technology, the complexity and scale of software systems are increasing day by day, and software tests play an increasingly important role as an important component of a life cycle. Engineering practices, however, have shown that software testing difficulty and cost are increasingly uncontrollable as software complexity and scale continue to scale. The traditional software testing method is obviously insufficient in the aspects of testing efficiency and fault location, and cannot meet the high-standard and high-requirement of modern software testing.
The traditional adaptive software test is based on a software control theory model and comprises tested software (controlled objects), a test strategy (controller) and a parameter estimation module. And (3) using the self-adaptive software test strategy collected on the test data line, estimating the required parameters, and selecting the next test case to form a software self-adaptive test theory. The conventional adaptive test model, as shown in FIG. 1, includes two feedback loops, the first of which occurs when a test decision is specified, and the next test case test strategy is generated from historical test data. A second feedback loop occurs in the historical test data to promote or alter the underlying test strategy. The main disadvantages of the traditional adaptive software test are that:
1) mode is difficult to fundamentally solve the problem of geometric increment of test time overhead
2) The algorithm efficiency problem under the condition of multi-dimension, the operation efficiency of the self-adaptive software test is not good under the condition of more than three dimensions, and a large amount of test overhead is consumed. The time complexity of the self-adaptive software test for screening a test case reaches O (n2), and the overhead of the algorithm for screening the test case influences the operation efficiency of the algorithm.
3) Under the condition of non-numerical value input field, the distance between the use cases is represented and calculated with great difficulty. The complexity and diversity of the test case types in the test process directly influence the application of the adaptive software test.
The method combines the software testing process to summarize the existing self-adaptive testing method, fully uses the existing software testing technology based on the software control theory idea for reference, analyzes the main problems existing in the software control theory model, points out the problems existing in the model, and then provides the view that the optimal testing decision has an unstable stage. Finally, aiming at the problems, an improved software self-adaptive testing method is provided by removing unreasonable assumptions in a model and introducing new assumptions, and engineering application shows that the self-adaptive testing method provided by the invention is remarkably improved in the aspect of testing efficiency and has higher practicability than a software cybernetics method.
Disclosure of Invention
The invention aims to provide an improved software self-adaptive testing method, which provides technical support for developing software self-adaptive testing with high efficiency.
The above object of the present invention is achieved by the following method: an improved self-adaptive test method is characterized in that a unique feedback control loop is formed by the test step length of an input domain, historical test data, parameter estimation and a controller test decision;
taking the tested input domain information as a test data history, and taking an adjustment algorithm as a parameter estimation module;
and selecting the partitions and the input domain as test decisions, and estimating the number of test cases of the input domain as the test step length design of the input domain.
Further, a test decision A is generated by the controllert+1And selecting a proper input field, and adjusting the test decision and the test step size of the input field based on the collected test data and the parameter estimation.
Further, the selecting the partition and the input field as a test decision specifically includes: after any test action is completed, there are 2 types of selectable test decisions, denoted as a ═ a0，a1In which a is0A child input field decision indicating selection of the current partition, a1Indicating a child input field decision to select a non-current partition.
Further, the specific steps of the adjustment algorithm of the adaptive test method are as follows:
(1) initializing an input domain set S of the tested software, and taking values from 1 to n for subscripts k of input domain partitions.
(2) The initial test decision and the probability of an in-input domain intra-partition transfer θ _0 and the probability of an out-of-input domain partition transfer θ _1 are set such that θ _0+ θ _1 is 1.
(3) The input fields (subscript k takes values 1 to n) are set as follows: setting the basic quantity L _ k of the test cases in the input domain partition C _ k to be C, setting the dynamic quantity DL _ k of the test cases in the input domain to be L _ k, and setting the upper limit a of the test process of DL _ k to be less than or equal to b (a, b and C are constants).
(4) One input domain partition C _ i (i is more than or equal to 1 and less than or equal to n) is randomly selected, and one input domain D _ j is selected from E _ i in the input domain partition C _ i.
(6) And testing by the number DL _ i of the test cases. The results were divided into 2 cases: if the defects are found, inputting the number F _ i +1 of the defects of the partition of the domain, and updating the test decision probability theta _0 to theta _0+ delta theta _ F and theta _1 to theta _ 1-delta theta _ F; if no defect is found, the test decision probability θ _0 ═ θ _0- Δ θ _ t, θ _1 ═ θ _1+ Δ θ _ t, (0 ≦ θ _0, θ _1 ≦ 1, if θ _0, θ _1<0, θ _0, θ _1 ═ 0, if θ _0, θ _1>1, θ _0, θ _1 ═ 1) is updated.
(7) Increment T _ i by 1 and update
If DL _ i is less than or equal to a, DL _ i is a; if DL _ i ≧ b, DL _ i ═ b.
(8) If the input field set E _ i-D _ j is not present
Step 12 is executed by making S-C _ i; otherwise, the next step is executed.
(9) If | S | ═ 1, θ _0 ═ 1, and θ _1 ═ 0. Step 11 is performed.
(10) If it is not
θ _0 equals 0, θ _1 equals 1; otherwise, the test process is finished and the test is quitted.
(11) Selecting an input domain partition in the test process according to the test decision probabilities theta _0 and theta _1, and selecting the input partition as a next test decision by using the probability theta _0, namely C _ i is equal to C _ i; selecting other input partitions with the probability of theta _1 as the next test decision, namely C _ i ═ C _ j (i ≠ j,1 ≦ j ≦ n, C _ j ∈ S), and updating theta _0, theta _1 according to the C _ i information (F _ i, T _ i). Step 4 is performed.
The invention has the beneficial effects that: according to key problems faced by software testing, the invention can effectively solve the problem of software input space combination explosion by improving the hypothesis of the traditional software self-adaptive testing method, fully utilizes intermediate testing data generated in the testing process, realizes dynamic and self-adaptive feedback regulation of the software testing process, can effectively reduce the number of test cases, improves the testing efficiency and reduces the testing cost. Therefore, the method has important engineering application value.
Drawings
FIG. 1 is a diagram of a conventional adaptive software test model.
FIG. 2 is a diagram of an improved software adaptive test model according to the present invention.
FIG. 3 is a flow chart of an improved software adaptive test algorithm of the present invention.
The specific implementation mode is as follows:
as shown in FIG. 2, the present invention provides an improved adaptive test method, which comprises a unique feedback control loop formed by the test step size of the input field, the historical test data, the parameter estimation and the test decision of the controller;
taking the tested input domain information as a test data history, and taking an adjustment algorithm as a parameter estimation module;
and selecting the partitions and the input domain as test decisions, and estimating the number of test cases of the input domain as the test step length design of the input domain.
Wherein the test decision A is generated by a controllert+1And selecting a proper input field, and adjusting the test decision and the test step size of the input field based on the collected test data and the parameter estimation.
The invention is based on the characteristics of the self-adaptive software model shown in figure 1 and the testing defects in engineering application, and carries out improvement and optimization to form an improved software self-adaptive testing method. In a traditional self-adaptive software test model, tested input domain information is used as test data (history), an adjustment algorithm is used as a parameter estimation module, a partition and an input domain are selected as test decisions, and the number of test cases in the input domain is estimated to be used as input domain test step length design.
The improved self-adaptive software testing method provided by the invention has the advantages that unreasonable assumptions in the software control theory model are removed, new assumption conditions are introduced, the self-adaptive random testing thought is combined, the software control theory model is improved by taking the tested software input field testing coverage information as a state transition condition, the testing resources are reasonably distributed by expanding the feedback link and improving the testing decision, and the testing efficiency is improved.
Compared with the traditional self-adaptive test model, the traditional self-adaptive test method selects the use case which is most likely to find the defects based on the current test information, and the idea of the improved software self-adaptive test model is to select a proper input domain and allocate proper test resources based on the existing test information.
Aiming at the improved software self-adaptive testing technology, the invention has the following specific implementation process:
one, the property that the effective adaptive test method should possess
Software self-adaptive testing based on a cybernetics model is abstract and hypothesis of a traditional software testing process, but partial hypothesis conditions of the model cannot truly reflect the actual situation of software testing. In current research and engineering practice, although a software control theory model is insufficient, a test method based on feedback control has the advantages of clear feedback link, acceptable complexity and the like, and the probability of finding software defects can be effectively improved by optimizing and improving a software test process.
On the other hand, as an important software self-adaptive testing method, the self-adaptive random testing method has the characteristics of simple and mature algorithm, and software defect aggregation property is fully utilized, so that the testing effect is better in a numerical input domain with lower dimensionality, but the method still has the defects of high computational complexity, poor applicability under the conditions of high dimensionality and non-numerical input domains and the like.
Based on the above analysis, an effective software adaptive test method should have at least the following properties:
the input domain of the tested software can be described;
the characteristic of software defect aggregation can be utilized;
intermediate test data generated in the test process are fully utilized;
dynamic and self-adaptive feedback regulation in the test process can be realized;
with acceptable computational complexity.
Second, optimization and improvement of software self-adaptive test mode
The following improvements and assumptions are made for the conventional adaptation test:
from the test coverage angle, the software test process is regarded as the process of completely testing all input domains in the software;
software defects have correlation and aggregation, and the number and the positions of the defects are unknown;
if a certain defect is detected, the defect is immediately removed, and a new defect is ensured not to be introduced;
the tested software is regarded as an input space consisting of a plurality of sub-input domains of different types, the input space is divided into n equivalent partitions, and the k equivalent partition comprises mkThe sub-input fields have similarity or correlation, and each equivalent partition is independent;
for n equivalent partitions, assume that only one of the following 2 states can be assumed:
set of available states of tested software states
Indicates that the initial time state of the test is
The state domain is discrete;
the test decision at time t is AtAnd performing a test action after the test decision: a complete test procedure involving a single sub-input domain;
whether the defects are found or not is all completed, the test case corresponding to the test step length is completed, (the test step length is dynamically adjusted along with the test), the test action is not stopped because the defects are found, each test can only be completed by one test action, the test process is performed in a single step, and the time domain is discrete;
the software defect found by each test action is marked as a defect, and a plurality of defects found in one test action are also marked as a defect;
after any test action is completed, there are 2 types of selectable test decisions, denoted as a ═ a0，a1In which a is0A child input field decision indicating selection of the current partition, a1A child input field decision representing a selection of a non-current partition;
the end state of the test is ξfinal＝[0，0，…，0]The test procedure is terminated.
Third, the improved software self-adaptive test algorithm of the invention
The core of the improved software self-adaptive testing algorithm is that tested input domain information is used as testing history data, an adjusting algorithm is used as a parameter estimation module, a partition and an input domain are selected as a next testing decision, the number of test cases in the input domain is estimated and used as an input domain testing step size design, and therefore testing efficiency is improved.
In the improved self-adaptive test model, the tested software, the test step length of the input domain, the historical test data, the parameter estimation and the test decision form a unique feedback control loop. As shown in FIG. 2, test decision A is generated by the controllert+1And selecting a proper input field, and adjusting the test decision and the test step size of the input field based on the collected test data and parameter estimation.
Suppose the measured softnessThe test case set of the device is regarded as the composition of the test case set partitioned by a plurality of input fields, and is divided into n equivalent classes which are marked as C1,C2,…CnEach equivalent input field partition comprises a plurality of input fields, which are marked as D1,D2,…,Dm. The tuning algorithm flow chart is shown in fig. 3, in conjunction with the model assumptions and the test model of fig. 2.
(1) Initializing an input domain set S of the tested software, and taking values from 1 to n for subscripts k of input domain partitions.
(2) The initial test decision and the probability of an in-input domain intra-partition transfer θ _0 and the probability of an out-of-input domain partition transfer θ _1 are set such that θ _0+ θ _1 is 1.
(3) The input fields (subscript k takes values 1 to n) are set as follows: setting the basic quantity L _ k of the test cases in the input domain partition C _ k to be C, setting the dynamic quantity DL _ k of the test cases in the input domain to be L _ k, and setting the upper limit a of the test process of DL _ k to be less than or equal to b (a, b and C are constants).
(4) One input domain partition C _ i (i is more than or equal to 1 and less than or equal to n) is randomly selected, and one input domain D _ j is selected from E _ i in the input domain partition C _ i.
(6) And testing by the number DL _ i of the test cases. The results were divided into 2 cases: if the defects are found, inputting the number F _ i +1 of the defects of the partition of the domain, and updating the test decision probability theta _0 to theta _0+ delta theta _ F and theta _1 to theta _ 1-delta theta _ F; if no defect is found, the test decision probability θ _0 ═ θ _0- Δ θ _ t, θ _1 ═ θ _1+ Δ θ _ t, (0 ≦ θ _0, θ _1 ≦ 1, if θ _0, θ _1<0, θ _0, θ _1 ═ 0, if θ _0, θ _1>1, θ _0, θ _1 ═ 1) is updated.
(7) Increment T _ i by 1 and update
If DL _ i is less than or equal to a, DL _ i is a; if DL _ i ≧ b, DL _ i ═ b.
(8) If the input field set E _ i-D _ j is not present
Step 12 is executed by making S-C _ i; otherwise, the next step is executed.
(9) If | S | ═ 1, θ _0 ═ 1, and θ _1 ═ 0. Step 11 is performed.
(10) If it is not
θ _0 equals 0, θ _1 equals 1; otherwise, the test process is finished and the test is quitted.
(11) Selecting an input domain partition in the test process according to the test decision probabilities theta _0 and theta _1, and selecting the input partition as a next test decision by using the probability theta _0, namely C _ i is equal to C _ i; selecting other input partitions with the probability of theta _1 as the next test decision, namely C _ i ═ C _ j (i ≠ j,1 ≦ j ≦ n, C _ j ∈ S), and updating theta _0, theta _1 according to the C _ i information (F _ i, T _ i). Step 4 is performed.
Claims (4)
1. An improved adaptive test method, characterized in that,
a unique feedback control loop is formed by the test step length of an input domain, historical test data, parameter estimation and a controller test decision;
taking the tested input domain information as a test data history, and taking an adjustment algorithm as a parameter estimation module;
and selecting the partitions and the input domain as test decisions, and estimating the number of test cases of the input domain as the test step length design of the input domain.
2. The improved software adaptive testing method of claim 1,
generating a test decision A by a controllert+1And selecting a proper input field, and adjusting the test decision and the test step size of the input field based on the collected test data and the parameter estimation.
3. The improved software adaptive testing method according to claim 1, wherein the selection of partitions and input fields as test decisions is specifically: after any test action is completed, there are 2 types of selectable test decisions, denoted as a ═ a0，a1In which a is0A child input field decision indicating selection of the current partition, a1Indicating a child input field decision to select a non-current partition.
4. The improved software adaptive test method as claimed in claim 1, wherein the adjustment algorithm of the adaptive test method comprises the following specific steps:
(1) initializing an input domain set S of the tested software, and taking values from 1 to n for subscripts k of input domain partitions;
(2) setting an initial test decision and an inside-input-domain-partition transition probability theta _0 and an outside-input-domain-partition transition probability theta _1 so that theta _0+ theta _1 is 1;
(3) the input fields (subscript k takes values 1 to n) are set as follows: setting the basic quantity L _ k of the test cases in the input domain partition C _ k to be C, setting the dynamic quantity DL _ k of the test cases in the input domain to be L _ k, and setting the upper limit a of the test process of DL _ k to be less than or equal to b (a, b and C are constants);
(4) randomly selecting an input domain partition C _ i (i is more than or equal to 1 and less than or equal to n), and selecting an input domain D _ j from E _ i in the input domain partition C _ i;
(6) the test is performed according to the number of test cases DL _ i, and the result is divided into 2 cases: if the defects are found, inputting the number F _ i +1 of the defects of the partition of the domain, and updating the test decision probability theta _0 to theta _0+ delta theta _ F and theta _1 to theta _ 1-delta theta _ F; if no defect is found, updating the test decision probability theta _0 to delta theta _ t, theta _1 to theta _1+ delta theta _ t, (0 to theta _0, and theta _1 to 1, if theta _0, theta _1 to 0, theta _0, and theta _1 to 0, if theta _0, theta _1>1, theta _0, and theta _1 to 1);
(7) increment T _ i by 1 and update
If DL _ i is less than or equal to a, DL _ i is a; if DL _ i is more than or equal to b, DL _ i is more than or equal to b;
(8) if the input field set E _ i-D _ j is not present
Step 12 is executed by making S-C _ i; otherwise, executeThe next step;
(9) if | S | ═ 1, θ _0 ═ 1, θ _1 ═ 0, step 11 is performed;
(10) if it is not
θ _0 equals 0, θ _1 equals 1; otherwise, the test process is finished and the test is quitted;
(11) selecting an input domain partition in the test process according to the test decision probabilities theta _0 and theta _1, and selecting the input partition as a next test decision by using the probability theta _0, namely C _ i is equal to C _ i; selecting other input partitions with the probability of theta _1 as the next test decision, namely C _ i ═ C _ j (i ≠ j,1 ≦ j ≦ n, C _ j ∈ S), updating theta _0 and theta _1 according to the C _ i information (F _ i, T _ i), and executing step 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010663807.9A CN112306859B (en) | 2020-07-10 | 2020-07-10 | Improved software self-adaptive testing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010663807.9A CN112306859B (en) | 2020-07-10 | 2020-07-10 | Improved software self-adaptive testing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112306859A true CN112306859A (en) | 2021-02-02 |
| CN112306859B CN112306859B (en) | 2022-07-08 |
Family
ID=74483174
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010663807.9A Active CN112306859B (en) | 2020-07-10 | 2020-07-10 | Improved software self-adaptive testing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112306859B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113111008A (en) * | 2021-05-12 | 2021-07-13 | 中国工商银行股份有限公司 | Test case generation method and device |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1770118A (en) * | 2004-10-27 | 2006-05-10 | 北京航空航天大学 | Adaptive test method for software |
| US20120049881A1 (en) * | 2010-08-30 | 2012-03-01 | International Business Machines Corporation | Integrated cross-tester analysis and real-time adaptive test |
| CN102831064A (en) * | 2012-09-04 | 2012-12-19 | 北京航空航天大学 | Software self-adaptive testing method facing to reliability assessment |
| CN103279422A (en) * | 2013-06-17 | 2013-09-04 | 东南大学 | Self-adaptive random test method based on rejecting area |
| US20160162392A1 (en) * | 2014-12-09 | 2016-06-09 | Ziheng Hu | Adaptive Framework Automatically Prioritizing Software Test Cases |
| CN106991039A (en) * | 2016-01-20 | 2017-07-28 | 滴滴（中国）科技有限公司 | Method of testing and device for platform adaptive automotive engine system |
| CN108804334A (en) * | 2018-06-15 | 2018-11-13 | 北京航空航天大学 | A kind of discrete type software reliability growth test and evaluation method based on adaptive sampling |
-
2020
- 2020-07-10 CN CN202010663807.9A patent/CN112306859B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1770118A (en) * | 2004-10-27 | 2006-05-10 | 北京航空航天大学 | Adaptive test method for software |
| US20120049881A1 (en) * | 2010-08-30 | 2012-03-01 | International Business Machines Corporation | Integrated cross-tester analysis and real-time adaptive test |
| CN102831064A (en) * | 2012-09-04 | 2012-12-19 | 北京航空航天大学 | Software self-adaptive testing method facing to reliability assessment |
| CN103279422A (en) * | 2013-06-17 | 2013-09-04 | 东南大学 | Self-adaptive random test method based on rejecting area |
| US20160162392A1 (en) * | 2014-12-09 | 2016-06-09 | Ziheng Hu | Adaptive Framework Automatically Prioritizing Software Test Cases |
| CN106991039A (en) * | 2016-01-20 | 2017-07-28 | 滴滴（中国）科技有限公司 | Method of testing and device for platform adaptive automotive engine system |
| CN108804334A (en) * | 2018-06-15 | 2018-11-13 | 北京航空航天大学 | A kind of discrete type software reliability growth test and evaluation method based on adaptive sampling |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113111008A (en) * | 2021-05-12 | 2021-07-13 | 中国工商银行股份有限公司 | Test case generation method and device |
| CN113111008B (en) * | 2021-05-12 | 2024-02-23 | 中国工商银行股份有限公司 | Test case generation method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112306859B (en) | 2022-07-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107506865B (en) | Load prediction method and system based on LSSVM optimization | |
| CN110110862A (en) | A kind of hyperparameter optimization method based on adaptability model | |
| CN112598150B (en) | Method for improving fire detection effect based on federal learning in intelligent power plant | |
| CN108762221B (en) | Deadlock free control method for automatic manufacturing system with uncontrollable event | |
| CN110837915B (en) | Low-voltage load point prediction and probability prediction method for power system based on hybrid integrated deep learning | |
| WO2022121518A1 (en) | Parameter configuration optimization method and system for distributed computing jobs | |
| CN109034479B (en) | Multi-target scheduling method and device based on differential evolution algorithm | |
| CN112306859B (en) | Improved software self-adaptive testing method | |
| CN113784410A (en) | Heterogeneous wireless network vertical switching method based on reinforcement learning TD3 algorithm | |
| CN111355633A (en) | Mobile phone internet traffic prediction method in competition venue based on PSO-DELM algorithm | |
| CN116402002A (en) | Multi-target layered reinforcement learning method for chip layout problem | |
| Zhang et al. | A simple yet universal strategy for online convex optimization | |
| CN117290721A (en) | Digital twin modeling method, device, equipment and medium | |
| CN115345297A (en) | Platform area sample generation method and system based on generation countermeasure network | |
| JP7137074B2 (en) | Optimization calculation method, optimization calculation device, and optimization calculation program | |
| CN114186709A (en) | Energy prediction method for optimizing key parameters of gray model based on emperor butterfly algorithm | |
| Zhou et al. | Imperialist competitive algorithm based on VNSOBL optimization for distributed parallel machine scheduling problem | |
| CN116149855A (en) | Method and system for optimizing performance resource cost under micro-service architecture | |
| CN116193516A (en) | Cost optimization method for efficient federation learning in Internet of things scene | |
| CN113157344B (en) | DRL-based energy consumption perception task unloading method in mobile edge computing environment | |
| GB2622756A (en) | Training agent neural networks through open-ended learning | |
| CN115129447A (en) | Method and computer equipment for unloading edge calculation tasks considering condition risk values in industrial internet | |
| CN111641974B (en) | Method and storage device based on 5G small-sized cellular hybrid renewable energy network | |
| Wei et al. | Fork: A forward-looking actor for model-free reinforcement learning | |
| CN114611897A (en) | Intelligent production line self-adaptive dynamic scheduling strategy selection method |
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 |