基于协同进化和并行计算的船舶管路布置方法

为解决船舶管路协同布置问题,提出一种适合求解多管路或分支管路协同布置的算法框架.通过为每条管路或分支生成对应的进化种群,将管路间的协同布置转换为种群间的协同进化.基于提出的路径连接点概念,生成管路接口间的候选路径种群,并对种群进行交叉、变异操作.使用A~*算法作为寻路算子,提高了生成路径的质量,同时保证了路径的有效性.为了提高运算效率,引入并行计算策略对算法框架和A~*算法进行改进.最后,两个仿真实例验证了所提出方法的可行性和有效性.

Method of ship pipe routing based on co-evolution and parallel computing

In order to solve the problem of cooperative pipe optimization in ship, an algorithm framework which is suitable for multiple pipes routing or branch pipe routing is presented. In this framework, evolutionary populations are created for the corresponding pipes or branches. Therefore, the cooperative arrangement of pipes or branches is transformed to the co-evolution of these populations. Based on the concept of connection points, the populations that represent the candidate paths between interface points are generated and gradually evolved through the operations like crossover and mutation. A * algorithm is used as the path finding operator, and the quality and validity of the paths are assured. To speed up the pipe routing, parallel computing technology is adopted to improve the proposed framework and A * algorithm. In the end, the feasibility and effectiveness of the proposed approach are verified by two simulation cases.