0-1 mathematical programming models for flexible process planning

Flexible Process Planning (FPP) is one of the key intelligent manufacturing techniques. The FPP problem is exactly and concisely formulated using 0-1 mathematical programming. Compared with the existing models, the new formulation simultaneously considers alternative operation selection and sequencing and operational method assignment under two optimization criteria. The new formulation does not need to plot the common AND/OR-network that often depicts partial possible processing routes. Distinctively, the important operational precedence constraint is beforehand transformed into the possible successor set of each operation and the possible immediate successor set. Three methods are creatively proposed to prohibit from generating a cycle in sequencing. The complicated criteria involving the machine, tool and setup changeover identification are linearly expressed. The experimental results indicate that the proposed 0-1 linear programming models are able to quickly obtain the optimal solution of the small-scale problems and stably find a satisfactory solution of the large-scale problems within acceptable time. Compared with the existing mathematical programming models for process planning, the proposed linear models have lower complexity and better performance in solving benchmark instances. In two groups of comparative experiments, the number of constraints of the proposed linear models dramatically reduces by 99.6% and 70%, respectively. Moreover, all benchmark instances are exactly solved by the Cplex solver using the proposed linear models within one hour.& COPY; 2022 Elsevier B.V. All rights reserved.

Automated summary

Flexible Process Planning (FPP) is a key intelligent manufacturing technique that is formulated using 0-1 mathematical programming. The new formulation simultaneously considers alternative operation selection and sequencing and operational method assignment under two optimization criteria. The proposed linear models have lower complexity and better performance in solving benchmark instances compared to existing mathematical programming models.