基于适应度值的信息反馈型群智能算法研究

项目名称： 基于适应度值的信息反馈型群智能算法研究

项目编号： No.61503165

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 其他

项目作者： 王改革

作者单位： 江苏师范大学

项目金额： 22万元

中文摘要： 本项目以适应度值对算法性能的影响为研究核心，与其他研究者关注的侧重点不同，本项目主要考虑如何根据适应度值初始化种群、动态调整参数及改进个体更新策略，旨在提高算法求解问题的能力和促进智能算法理论的发展。主要内容包括：(1) 通过均匀分割的确定性和莱维分布的不确定性来初始化种群；(2) 分析种群适应度值的统计特征，提出种群优劣评价方法；(3) 根据种群优劣评价方法判断种群优劣，结合混沌理论对算法参数进行动态调整；(4) 将适应度值信息反馈到算法个体更新过程，对个体更新策略进行改进。在应用方面，拟将取得研究成果应用于解决系统可靠性、调度、背包和任务分配等问题。

中文关键词： 群体智能优化算法；混合群体智能优化；算法设计与实现

英文摘要： The core of the project is the impacts of fitness values on the performance of the algorithm, which is different from other researchers’ focus. The project mainly tends to investigate how to initialize population, dynamically adjust parameters and improve the individual updating process according to fitness values, with the aim of improving the ability of the algorithm for solving problem and promoting the development of intelligent algorithm theory. A novel information feedback type swarm intelligence algorithm based on the characteristics of fitness values will be proposed in this project. The main contents can be described as follows. (1) Population is initialized by combining the determinateness of uniform segmentation and the indeterminateness of Lévy distribution. (2) By analyzing the statistical characteristics of fitness values, the evaluation of population is put forward. (3) Judging the population quality by the population evaluation method, the parameters are dynamically adjusted by the combination of chaos theory. (4) The individual updating process will get feedback from fitness values for each generation, and accordingly it is then improved. In the aspect of application, the obtained research results in this project intend to apply to solve the reliability of the system, scheduling, knapsack and task assignment.

英文关键词： swarm intelligence optimization algorithm;Hybrid swarm intelligence optimization ;Algorithm design and implementation