基于多目标遗传算法的RITS物理结构优化设计

铁路智能运输系统(RITS)是涉及多学科、多专业领域,集底层控制、实时调度、运营管理于一体的多功能、多任务的复杂大型信息系统,系统结构优化设计方法研究对铁路信息化建设和现有信息化资源的整合具有重要意义。本文提出了RITS物理结构优化设计问题模型,即在一定资源条件约束下,综合考虑技术、经济等方面因素,实现功能单元在物理实体上的优化配置,并提出了基于多目标优化的解决思路。RITS结构优化设计待优化函数具有高维、多峰、非线性等特点,求解难度大。本文重点研究了一类基于Pareto机制的具有快速全局收敛能力的多目标遗传算法,并以紧急救援系统为例进行结构优化设计,验证了算法的有效性。

Optimized Design of RITS Physical Structure Based on Multi-objective GA

As a large distributed information system,Railway Intelligent Transportation Systems(RITS) involves modern control,real-time scheduling,business operation management and such theoretical fields to provide different services for physical distributed customers or designers.To feed these multi-service and multi-task requirments,the methodology for the system structure design plays a most critical role in the development of RITS.An optimal design model for the physical structure of RITS has been presented,which actualizes the physical structure optimization after considering the factors of technology and economy.On the base of characteristics and difficulties of RITS melioration design and structure optimization,multi-objective is used to meet design goals utilizing system resources in an effective way.The pre-optimized function for RITS structure design features with high dimension,multi-peak,non-linearity and computational complexity and there is usually no optimal unique solution in traditional meanings.A MOGA based Pareto optimization approach is put forward with its effectiveness and efficiency approved in the design of the railway emergency system.