首页> 外国专利> DISCRETE DIFFERENTIAL EVOLUTION-BASED MULTI-STRATEGY OPTIMIZATION X-ARCHITECTURE MINIMUM TREE CONSTRUCTION METHOD

DISCRETE DIFFERENTIAL EVOLUTION-BASED MULTI-STRATEGY OPTIMIZATION X-ARCHITECTURE MINIMUM TREE CONSTRUCTION METHOD

机译:基于离散的差分演进的多策略优化X架构最小树施工方法

摘要

A discrete differential evolution-based multi-strategy optimization X-architecture minimum tree construction method, comprising the following steps: step S1, reading pin information of a circuit to be tested; step S2, initializing a population, calculating an individual adaptation value of the initial population, and initializing an adaptive parameter; step S3, determining whether the number of algorithm iterations reaches a threshold; step S4, if the threshold is not reached, randomly selecting a variation strategy from a strategy pool, and obtaining a child individual after variation and crossing; if the threshold value is reached, carrying out variation and crossing operations on the population according to a conventional differential evolution algorithm; step S5, adopting an immune clone selection strategy; step S6, determining whether iteration meets a termination condition, if yes, terminating iteration and outputting a final population, and otherwise, returning to step S3 to continue the next iteration update; and step S7, adopting a refining strategy to obtain an optimal scheme of a wiring tree. According to the method, a larger search space and a smaller wire length can be obtained, a topological structure of the wiring tree is optimized, and redundancy is reduced.
机译:基于离散的基于差分演化的多策略优化X架构最小树施工方法,包括以下步骤:步骤S1,读取要测试的电路的引脚信息;步骤S2,初始化群体,计算初始群体的单独适应值,并初始化自适应参数;步骤S3,确定算法迭代的数量是否达到阈值;步骤S4,如果未达到阈值,则随机选择从策略池中的变化策略,并在变化和交叉后获得子个体;如果达到阈值,则根据传统的差分演进算法对群体进行变化和交叉操作;步骤S5,采用免疫克隆选择策略;步骤S6,确定是否迭代是符合终止条件,如果是,则终止迭代并输出最终群体,否则返回步骤S3以继续下次迭代更新;和步骤S7,采用精制策略来获得布线树的最佳方案。根据该方法,可以获得更大的搜索空间和较小的线长度,优化布线树的拓扑结构,并且减少了冗余。

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号