基于生物网络的高维多目标算法及其在分布式调度中的应用

项目名称： 基于生物网络的高维多目标算法及其在分布式调度中的应用

项目编号： No.71471158

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 管理科学

项目作者： 陈东燊

作者单位： 香港理工大学深圳研究院

项目金额： 62万元

中文摘要： 设计启发式算法求解带有不确定参数的高维多目标优化问题是管理科学、运筹学领域中一个研究热点和难点。本项目借鉴生命= DNA+环境+环境与遗传交互作用的思想，对个体生命进化行为建模与仿真，将个体的进化行为置于细胞→个体→环境→细胞构成的生物网络与社会网络的结点中，研究个体的进化过程，进而提取一种基于生命进化行为的新型生物网络优化算法(BSNO)并开展在分布式调度中的应用。具体包括：①高维多目标算法统一模型构建和算法的统一实验仿真平台的构建；②基于个体生命进化行为建模与仿真；③ 基于生命进化行为的生物社会网络算法的提出及理论分析和性能测试；④整合机器维护和分布式调度系统的生产联合决策模型的建立（JDMM)；⑤ 设计BSNO算法优化JDMM模型框架及算法流程。该研究不仅丰富了多目标优化理论，而且为带有不确定因素的高维多目标分布式调度的求解提供了新思路，具有重要的科学意义和应用价值。

中文关键词： 生物网络；多目标优化；算法；分布式调度

英文摘要： In the field of management science and operations research, development of heuristic algorithms to deal with high-dimensional multi-objective optimization problems with uncertain parameters is a high demand and complex research. The idea of this project is stimulated from the concept of 'Life = DNA+environment+interaction of environment and genetic DNA'. Modeling and simulation of individual life evolution behavior will be conducted. Furthermore, the evolution of individual behavior in the biological and social network nodes composed of 'cells-individual-environment-cells' will be studied. As a result, a new biological network optimization algorithm based on evolutionary behavior will be developed.The entire project will be sub-divided into five phases or work packages including (1) building high-dimensional multi-objective unified model and algorithm simulation platform; (2) developing computer modeling and simulation based on individual life evolution behavior; (3) proposing Biological Social Network Optimization (BSNO) algorithm and conducting the theoretical analysis and performance testing on the proposd BSNO; (4) integrating machine maintenance and distributed scheduling system of the proposed production Joint Decision Making Model (JDMM); and (5) designing BSNO framework and process to optimize the JDMM model. This research will not only enrich the theory of multi-objective optimization, indeed a new method of optimizing multi-objective distributed scheduling problems will be developed, which has significant scientific merits and great potenatil of real-life industrial applications.

英文关键词： biological network;multi-objective optimization;algorithm;distributed scheduling