Research on repair operators in the whole space search genetic algorithm of assembly job shop scheduling problem

There are multistage constraints among workpieces in AJSSP (Assembly Job Shop Scheduling Problem) and the infeasible working procedures to assembly restraints are needed to be decoupled into feasible ones in which the sequences satisfy assembly relations (corresponding to the chromosomes in GA, which refers to Genetic Algorithm), before distributing to machines. Therefore the difficulty of whole space search genetic algorithm to AJSSP lies in the adjustment from infeasible chromosomes to feasible ones. In this paper a repair operator of TDRA (Top-Down Recursively Adjustment) is proposed, which means recursive post adjustment of forefather working procedures to an assembly tree structure step by step from the whole tree family to small branch families, in order to eliminate all the violating restraints and repair the infeasible chromosomes to feasible ones. In addition the evolution effects and optimization searching characteristics applying different repair operators are compared via experiments, and the results demonstrate that the evolution effect is obvious when applying TDRA to repair GA in AJSSP and optimal solutions can be searched out each time.