An Auto-MILP Model for Flexible Job Shop Scheduling Problem
Liping Huang, Rong Su
Abstract
An auto mixed integer linear programming (Auto-MILP) model is proposed to tackle the flexible job shop scheduling problem. The Auto-MILP model allows the precedence between operations of a job to be given by an arbitrary directed acyclic graph rather than a linear order. The goal is the minimization of the makespan. By incorporating the operation assignment to resources, the Auto-MILP model also allows multiple production lines in the workshop, which means more than one machine could complete the same operation, and the machine assignment is determined by the algorithm. We utilize time separation to define the machine capacity, which makes it adaptive to various kinds of job scheduling scenarios. The effectiveness and efficiency of the Auto-MILP model are evaluated using instances from simulated scheduling case and real data from an industrial job shop.