5.3 Comparative study with simulated annealing (SA)
SA is a general stochastic computational method for finding global extremums to large optimization problems, which can also be used in the concurrent optimization of process plan design and configurations of RMT[16]. We also implemented SA to optimize the problem. The optimization problem can be formulated as a pair of (p, f ), where p is similar to chromosome of GA which is composed of three parts. f is the objective function that is also the same. The computation pseudo code is as follows: procedure simulated repeat repeat select a new p; if f(p)f(p) then pp; else if random[0, 1]exp((f(p)f(p))T) then pp; until (termination-condition) T decrease 10%; Until (stop-criterion) End; The feasible set is generated from OP graph with the same method as GA. In our implementation, SA obtained the same optimization value as GA. However, the convergence rate is much slower than GA which cost 35s. It may relate with the parameter selection of SA which is dependent on experience to a great extent. Since it is well known that SA helps to avoid local minimum solution, the combination of GA and SA should be better solution which is also our work in the future.
摘要:生产过程计划设计和配置的可重构机床(RMT)相互作用。RMT合理的工艺计划与合适的配置,有助于提高产品质量和降低生产成本。因此,合作策略是需要同时解决上述问题。合作优化模型配置和生产过程计划RMT提出。其目标考虑的过程和配置都影响。此外,遗传算法是一种新颖的开发提供最优或接近最优的解决方案:首先,其染色体是重新设计的,它由三个部分组成,业务,流程规划和配置,分别RMT;其次,它的新选择、交叉和变异算子也发展到处理过程约束从操作过程,否则这些运营商可能产生非法的解决方案违反限制;最终的最优配置方案设计优化过程下RMT可以获得。最后,生产线案例应用的三RMT组成。它显示的情况下,最优过程计划和配置的同时获得RMT,生产成本降低6.28%,非货币的性能提高了22%。该方法可以找出两个RMT配置和生产过程,提高生产能力,功能和设备利用率。
关键词:可重构制造系统,可重构机床、配置、流程计划,合作优化模型