基于蒙特卡洛仿真和并行粒子群优化算法的携行备件优化

Optimization of Carried Spare Parts Based on Monte Carlo Simulation and Parallel Particle Swarm Optimization Algorithm

为提高携行备件方案优化模型的准确性和求解的精确度,以遂行远海训练任务的舰艇编队为研究背景,针对优化模型的建立和求解提出了一系列改进措施。在传统优化模型的基础上,分析了虚警和串件拼修对备件的影响,建立了基于携行能力、备件成本、装备可用度、同型号装备群完好率等多约束条件的携行备件优化模型;利用粒子群优化(PSO)算法确定备件的优化配置,利用蒙特卡洛仿真法计算配置方案的保障效能;引入云格计算技术实现PSO算法的并行求解,从硬件性能上提高算法的全局寻优能力;将普通粒子转化为量子粒子实现解的多样化,减小了算法陷入局部最优的危险。案例分析证实了改进措施的可行性和有效性。

In order to enhance the veracity of model and the accuracy of solution for carried spare parts scheme, several improvements are raised under the background of naval fleet. An improved model is established, which is based on multi-constraints including carrying capacity, spare parts cost, availability of equipment, and serviceability rate of the same type equipment. Particle swarm optimization（PSO） is used to optimize the spare parts scheme, and Monte Carlo simulation is used to calculate its performance. PSO is concurrently computed on the designed gloud platform, which could improve global optimization a- bility by hardware parallel computation, and the ordinary panicles in PSO are transformed to the quantum particles in special conditions, which could avoid PSO trapping in local optimum. The example shows that these improvements are feasible and effective.