仿真优化中资源分配问题的研究物流与运营管理专业论文.docxVIP

摘要当前，系统仿真已经成为任何复杂系统，特别是高技术产业不可缺少的分析、 摘要 当前，系统仿真已经成为任何复杂系统，特别是高技术产业不可缺少的分析、 研究、设计、评价、决策和训练的重要手段，其应用范围在不断扩大，应用效益 也日益显著。仿真优化对系统仿真的效果尤其重要，一般将仿真优化方法分为传 统仿真优化方法和仿真资源分配方法。 系统仿真最大的难点在于计算过于费时。因此，将有限的计算资源分配到需 要的方案上尤为重要。最优计算预算分配方法，即oCBA(optiIllal Computing Budget触location)方法，正是以此为目标，通过科学的计算，将计算资源合理 分配到各个方案，增强仿真的效率和有效性。 目前已有很多关于oCBA方法的研究，但都未曾考虑仿真中不同方案的计算可 能具有相同的过程，所以多个方案可以在同一个计算过程中运算，以节约计算资 源。本文在之前研究模型的基础上，假设有一种更有效的方法来估计各个方案， 即不同方案之间使用相同的循环算法，该新问题具有更好的通用性和实用性。对 该问题进行建模，最终可化简为线性规划问题，使用CPLEX等软件可以计算结果。 同时，还可以构建一种启发式算法进行计算。实际算例中发现，该问题的求解效 果比原模型有大幅提升。 关键词：仿真，优化，分配，0CBA 中图分类号：C935管理计划和控制 AbstractNowadays， Abstract Nowadays， simulation has become quite important for any complex system， especially the analysis， research， design， evaluation， decision and training of high—tech industries． The scope of its application is expanding constantly and its efficiency is increasing significantly． Simulation optimization is particularly important for the effectiVeness of simulation． Generally speaking， simulation optimization methods can be divided into the two parts， which are traditional methods and simulation resource allocation methods respectively． For simulation， the greatest difficulty is time—consuming． Therefore， determining the best allocation of simulation replications among Various designs becomes particularly important． The optimal Computing Budget A110cation(0CBA)method is to optimally allocate the budget to different designs through scientific calculation and eVentually improVe the efficiency and effectiveness of simulation． There have been a 10t of research about 0CE媳， however， no one cons idered that different designs may share the same calculation process to saye resource．On the basis of previous research， we establ ish a new model supposing that there is a more effectiVe way to estimate designs．The new model is more general and practical． It eVentually simplifies to a 1inear programming problem and we can solVe it by taking calculation software， such as CPLEX．Meanwhile，we al