基于蜂群算法和多智能体的多目标空间位置优化搜索和并行计算研究

Optimal Decision of spatial location is a complex and non-structured problem that human society usually confronted. Multi-objective spatial location optimization, for its objective functions' incommensurate and conflicting essence, has become a difficult problem. Optimal search of spatial location is a key technique to solve spatial location optimization problem, and its further study has important influence on promoting the ability of GIS spatial decision supporting and society resource optimization. Many traditional GIS optimal approaches reach their limits while searching optimal solution in multi-objective spatial location problems. Bee Colony Algorithm(ABC) is a kind of new swarm intelligent algorithms that performs well in global optimization. By importing the method of Non-dominated sorting， a new Pareto-ABC intelligent model can be proposed, which can effectively solve multi-objective spatial location optimization problem with the participation of huge amount of GIS data. Meanwhile, by introducing multi-Agent intelligent model and adopting distributed deployment strategies, a parallel computing environment can be constructed with ABC, which can improve problem-solving efficiency significantly. This study will make a useful exploration in promoting the GIS application in optimal decision of spatial location, which has significant scientific meanings and application values.

空间位置优化决策是人类社会经常面临的一个复杂非结构化问题，多目标空间位置优化问题，因多个目标内容之间的不可公度性和矛盾性，更成为GIS实际应用难题。空间位置优化搜索是求解空间位置优化问题的关键技术，该技术的深入研究对增强GIS空间决策支持能力和优化社会资源等具有重要意义。诸多传统的GIS优化方法在多目标空间位置优化问题寻求最优解的过程中体现出较大的局限性。蜂群算法作为在全局寻优上有着突出表现的一类"年轻的"群智能算法，通过引入Pareto的快速非支配选择排序法，构建新型的Pareto-蜂群优化模型，能有效地解决海量GIS数据参与下的多目标空间位置优化决策问题。同时，引入多Agent智能模型，采用分布式部署策略，构建一种结合蜂群优化的群智能并行计算环境，从而大幅提升问题求解效率。该研究项目的实施能为进一步推进GIS深入应用于空间位置优化决策做出有益的探索，具有重要的科学意义和应用价值。

空间位置优化决策是人类社会经常面临的一个复杂非结构化问题,多目标空间位置优化问题,因多个目标内容之间的不可公度性和矛盾性,更成为 GIS 实际应用难题。空间位置优化搜索是求解空间位置优化问题的关键技术,该技术的深入研究对增强 GIS 空间决策支持能力和优化社会资源等具有重要意义。诸多传统的 GIS 优化方法在多目标空间位置优化问题寻求最优解的过程中体现出较大的局限性。蜂群算法作为在全局寻优上有着突出表现的一类"年轻的"群智能算法,通过引入 Pareto 方法,构建新型的 Pareto-蜂群优化模型,能有效地解决海量 GIS 数据参与下的多目标空间位置优化决策问题。本项目以城市土地利用优化配置和城市生态控制线优化为应用点，构建了基于人工蜂群算法的多目标优化模型。同时,引入多 Agent 智能模型,采用分布式部署策略,构建一种结合蜂群优 化的群智能并行计算环境,从而大幅提升问题求解效率。该研究项目的实施能为进一步推进 GIS 深入 应用于空间位置优化决策做出有益的探索,具有重要的科学意义和应用价值。