Volume 22, Issue 2 (9-2025)
JSDP 2025, 22(2): 3-30 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Sadeghi Moghadam M, nejatian S, Parvin H, Bagheri Fard K, Yagoubian S H. Solving dynamic optimization problems with an Improved Imperialis Competition Algorithm. JSDP 2025; 22 (2) : 1
URL: http://jsdp.rcisp.ac.ir/article-1-1394-en.html
URL: http://jsdp.rcisp.ac.ir/article-1-1394-en.html
Mahdi Sadeghi Moghadam 
, Samad Nejatian , Hamid Parvin * , Karamullah Bagheri Fard , Seyed Hadi Yagoubian
, Samad Nejatian , Hamid Parvin * , Karamullah Bagheri Fard , Seyed Hadi Yagoubian
Islamic Azad University, Noorabad Mamsani, Fars, Iran
Abstract: (24 Views)
Many practical problems in optimization are considered static, however, many real-world problems are dynamic. Dynamic problems, along with nonlinear constraints and different objectives, are one of the features that have emerged in real world problems. Therefore, solving dynamic optimization problems has become one of the most important issues in the real world. Due to the fact that dynamic problems are NP-hard, an algorithm capable of solving such optimization problems should be provided. Therefore, in order for the algorithm to be able to optimally track the variable in these problems, it is necessary to provide an algorithm that can show proper efficiency in facing dynamic environments. There are particular challenges to optimization in dynamic environments: information about the problem propagates over time, uncertain events may occur, or the requirements of the problem may change over time. A common way to solve dynamic optimization problems is to use evolutionary algorithms. Today, the use of evolutionary algorithms based on natural biological behaviors is of particular importance to solve these problems. Among these algorithms, we can mention the Imperialist Competition optimization Algorithm (ICA). The purpose of this paper and its main claim is the possibility of designing protocols inspired by nature in the ICA algorithm, which is effective on optimization in dynamic environments, while maintaining the complexity of the algorithm and the changes in the problem space occur periodically. Among the various optimization algorithms based on swarm intelligence, ICA has attracted a lot of attention. Several studies have used ICA in static optimization and the results indicate good performance in static environments and poor performance in dynamic environments. In this paper, a new optimization algorithm in dynamic environments is proposed based on the ICA approach, in which various mechanisms are used to face the challenges of this field. In this algorithm, the improved multi-swarm approach is used to find the optimum (optimums) in the problem space and track them after changing the environment at the right time. The multi-swarm solution can divide the population into two sub-swarm in the problem space. Multi- swarm, instead of a single swarm, can increase the diversity in the problem space, and finally the convergence of the population particles to the optimum (optimums) takes place faster. The problem that exists in most multi- swarm methods is that the speed and efficiency of the algorithm gradually decreases with the uncontrolled increase of the population. The multi- swarm method presented in this paper is adaptive to the problem space, and whenever there is a need to increase the population, a population is created adaptively, and this makes the problem of the previous methods to be promoted. This modified ICA algorithm is based on memory and clustering. If the changes occur periodically, usually using the past information allows the algorithm to quickly adapt to the new environmental conditions after the environment changes. The desired idea in this field is to use a memory. One of the issues that are used today to solve problems, including those that face uncertainty, is that by having information from the recent past, one can predict the near future. In this article, according to this issue, a special type of memory is used to preserve past information. In this method, a new memory is used. This memory solves the defects of the standard memory and improves the efficiency of the proposed algorithm. This modified ICA algorithm is based on memory and clustering. Combining the proposed memory with clustering has been able to increase the performance of the proposed method in tracking change optima by creating enough diversity in each group of the population. This paper uses a new type of memory that creates a compromise between exploration and exploitation in the proposed algorithm. Using the appropriate clustering method can increase the diversity during the execution time of the algorithm. In this method, the k-means clustering technique is used in memory to maintain the diversity level. The proposed method has been tested on a benchmark called the second Brunk scenario called MPB. MPB is a suitable measure for evaluating the performance of optimization algorithms in dynamic environments and can also be used for very large-scale optimization problems. The proposed algorithm is compared with FTmPSO (TMO), RAmQSO-s4, RmNAFSA-s4, TFTmPSO, RFTmPSO, mQSO10 (5+5q), FMSO, CellularPSO, Multi-SwarmPSO, mCPSO, *AmQSO, FTMPSO, almPSO, and CDEPSA. The experimental results show that the proposed method has better efficiency than State-of-the-Art methods.
Article number: 1
Keywords: Dynamic Optimization, Dynamic Environments, Memory, Imperialist Competition Algorithm, Clustering, Moving Peaks Benchmark.
Type of Study: Research |
Subject:
Paper
Received: 2023/09/1 | Accepted: 2025/07/21 | Published: 2025/09/13 | ePublished: 2025/09/13
Received: 2023/09/1 | Accepted: 2025/07/21 | Published: 2025/09/13 | ePublished: 2025/09/13
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
