Parameter settings.

Differential Evolution (DE) is widely recognized as a highly effective evolutionary algorithm for global optimization. It has proven its efficacy in tackling diverse problems across various fields and real-world applications. DE boasts several advantages, such as ease of implementation, reliability, speed, and adaptability. However, DE does have certain limitations, such as suboptimal solution exploitation and challenging parameter tuning. To address these challenges, this research paper introduces a novel algorithm called Enhanced Binary JADE (EBJADE), which combines differential evolution with multi-population and elites regeneration. The primary innovation of this paper lies in the introduction of strategy with enhanced exploitation capabilities. This strategy is based on utilizing the sorting of three vectors from the current generation to perturb the target vector. By introducing directional differences, guiding the search towards improved solutions. Additionally, this study adopts a multi-population method with a rewarding subpopulation to dynamically adjust the allocation of two different mutation strategies. Finally, the paper incorporates the sampling concept of elite individuals from the Estimation of Distribution Algorithm (EDA) to regenerate new solutions through the selection process in DE. Experimental results, using the CEC2014 benchmark tests, demonstrate the strong competitiveness and superior performance of the proposed algorithm.

差分进化算法（Differential Evolution, DE）是一种被广泛认可的高效全局优化进化算法，已在跨多领域的各类问题及实际应用场景中验证了其有效性。DE具备诸多显著优势，诸如易于实现、可靠性强、运算速度快且适应性佳。但该算法仍存在一定局限：一方面局部开发能力不足，易生成次优解；另一方面参数调优难度较大。为解决上述问题，本研究提出一种名为增强型二进制JADE（Enhanced Binary JADE, EBJADE）的新型算法，该算法将差分进化与多种群策略及精英再生机制相结合。本文的核心创新之处在于引入一种具备更强局部开发能力的搜索策略：该策略基于对当前代三个向量的排序结果对目标向量进行扰动，并引入方向差分，引导搜索过程向更优解方向推进。此外，本研究采用带有奖励型子种群的多种群策略，动态调整两种不同变异策略的资源分配比例。最后，本文引入分布估计算法（Estimation of Distribution Algorithm, EDA）中的精英个体采样思路，通过差分进化的选择流程生成新的候选解。基于CEC2014基准测试集的实验结果表明，所提算法具备极强的竞争力与更优异的整体性能。