|
چکیده
|
In previous investigations in the eld of exible ow shop scheduling problems, the rework probability for operations was ignored. As these kinds of problems are NP-hard, we presented an Enhanced Invasive Weed Optimization (EIWO) meta-heuristic algorithm, in order to solve the addressed problem, with probable rework times, transportation times with a conveyor between two subsequent stages, di erent ready times and anticipatory sequence dependent setup times. The optimization criterion is to minimize makespan. Although Invasive Weed Optimization (IWO) is an ecient algorithm and has been used by many researchers recently, to increase the capability of IWO, we added a mutation operation to enhance the exploration in order to prevent sticking in local optimum. In addition, an anity function is embedded to obstruct premature convergence. With these changes, we balance the exploration and exploitation of IWO. Since the performance of our proposed algorithm depends on parameters values, we applied the popular design of an experimental methodology called the Response Surface Method (RSM). To evaluate the proposed algorithm, rst, some random test problems were generated and compared with three benchmark algorithms. The related results were analyzed by statistical tools. The experimental results and statistical analyses demonstrated that the proposed EIWO was e ective for the problem.
|