基于多群体的变异策略集成的差分进化

1. 题目与关键词

   • 英文题目：Differential evolution with multi-population based ensemble of mutation strategies

   • 英文关键词：Evolutionary algorithm, Differential evolution, Multi-population, Ensemble of mutation strategies, Numerical optimization

   • 中文题目：基于多群体的变异策略集成的差分进化

   • 中文关键词：进化算法，差分进化，多群体，变异策略集成，数值优化

2. 摘要

   • 英文摘要：
     Differential evolution (DE) is among the most efficient evolutionary algorithms (EAs) for global optimization and now widely applied to solve diverse real-world applications. As the most appropriate configuration of DE to efficiently solve different optimization problems can be significantly different, an appropriate combination of multiple strategies into one DE variant attracts increasing attention recently. In this study, we propose a multi-population based approach to realize an ensemble of multiple strategies, thereby resulting in a new DE variant named multi-population ensemble DE (MPEDE) which simultaneously consists of three mutation strategies, i.e., “current-to-pbest/1” and “current-to-rand/1” and “rand/1”. There are three equally sized smaller indicator subpopulations and one much larger reward subpopulation. Each constituent mutation strategy has one indicator subpopulation. After every certain number of generations, the current best performing mutation strategy will be determined according to the ratios between fitness improvements and consumed function evaluations. Then the reward subpopulation will be allocated to the determined best performing mutation strategy dynamically. As a result, better mutation strategies obtain more computational resources in an adaptive manner during the evolution. The control parameters of each mutation strategy are adapted independently as well. Extensive experiments on the suit of CEC 2005 benchmark functions and comprehensive comparisons with several other efficient DE variants show the competitive performance of the proposed MPEDE.

   • 中文摘要：
     差分进化（DE）是用于全局优化的最有效进化算法之一，目前广泛应用于解决各种实际应用问题。由于DE的不同配置可能显著不同，因此将多种策略适当地组合到一个DE变体中，以有效解决不同的优化问题，最近引起了越来越多的关注。在本研究中，我们提出了一种基于多群体的方法来实现多种策略的集成，从而产生了一种新的DE变体，名为多群体集成DE（MPEDE），它同时包括三种变异策略，即“current-to-pbest/1”、“current-to-rand/1”和“rand/1”。存在三个等大小的较小指示子群体和一个更大的奖励子群体。每种构成的变异策略都有一个指示子群体。在每一定数量的代之后，将根据适应度改进与消耗的函数评估之间的比例，确定当前表现最佳的变异策略。然后，奖励子群体将动态分配给确定的最佳表现变异策略。结果，更好的变异策略在进化过程中以自适应的方式获得更多的计算资源。每种变异策略的控制参数也是独立适应的。在CEC 2005基准测试函数集上的广泛实验和与几种其他高效DE变体的全面比较，显示了所提出的MPEDE的竞争性能。

3. 创新点与学术价值

   • 原始创新点：
     The study introduces a novel DE variant named MPEDE, which utilizes a multi-population based approach to dynamically combine multiple mutation strategies into one variant. It includes three mutation strategies: “current-to-pbest/1”, “current-to-rand/1”, and “rand/1”. The algorithm dynamically allocates a larger reward subpopulation to the best performing mutation strategy, allowing for adaptive computational resource allocation during evolution.

   • 中文创新点：
     本研究引入了一种名为MPEDE的新DE变体，它利用基于多群体的方法将多种变异策略动态地结合到一个变体中。它包括三种变异策略：“current-to-pbest/1”、“current-to-rand/1”和“rand/1”。算法动态地将较大的奖励子群体分配给表现最佳的变异策略，允许在进化过程中进行自适应的计算资源分配。

   • 学术价值：
     MPEDE的提出为差分进化算法提供了一种新的策略集成框架，能够自适应地调整计算资源分配，这有助于提高算法解决复杂优化问题的效率和效果。通过动态调整种群结构和参数设置，MPEDE能够更好地平衡探索和开发，增强算法的全局搜索能力和局部搜索精度，这对于进化算法领域是一个重要的进步。

4. 结论

   • 原始英文结论：
     Mutation strategies can significantly affect the performance of differential evolution (DE) on different optimization problems. In addition, different mutation strategies with different advantages can complement one another even when DE is applied to one optimization problem. As a result, an efficient integration of different mutation strategies is a promising way to enhance the performance of DE. In this paper, a multi-population based approach is proposed to realize the adapted ensemble of three mutation strategies into a novel DE variant, i.e. MPEDE. The population of MPEDE is dynamically partitioned into several subpopulations including three indicator subpopulations and one reward subpopulation as the algorithm proceeds. Each indicator subpopulation with relatively smaller size is assigned to a constituent mutation strategy and the reward subpopulation with relatively larger size is assigned to the currently best performed mutation strategy as an extra reward. Through this manner, we effectively realize the dynamic computation resource allocation among mutation strategies and the best mutation strategy is expected to timely obtain most computational resources (given with the reward subpopulation). In addition to the adaptive ensemble of multiple mutation strategies, the control parameters of each mutation strategy are adapted independently. Extensive experiments on CEC 2005 benchmark suit show that MPEDE outperforms several other efficient and popular peer DE variants including JADE, jDE, SaDE, EPSDE, CoDE, and SHADE.

   • 中文结论：
     变异策略可以显著影响差分进化（DE）在不同优化问题上的性能。此外，即使DE应用于一个优化问题，具有不同优势的不同变异策略也可以相互补充。因此，有效地整合不同的变异策略是提高DE性能的有希望的途径。在本文中，提出了一种基于多群体的方法，将三种变异策略适应集成到一种新的DE变体中，即MPEDE。MPEDE的种群动态划分为几个子群体，包括三个指示子群体和一个奖励子群体。每个较小的指示子群体分配给一个构成的变异策略，而较大的奖励子群体分配给当前表现最佳的变异策略作为额外的奖励。通过这种方式，我们有效地实现了变异策略之间的动态计算资源分配，最佳变异策略有望及时获得大多数计算资源（由奖励子群体提供）。除了适应性地集成多种变异策略外，每种变异策略的控制参数也是独立适应的。在CEC 2005基准测试上的广泛实验表明，MPEDE优于其他几种高效且流行的同类DE变体，包括JADE、jDE、SaDE、EPSDE、CoDE和SHADE。