Intelligent Scheduling for a Rolling Process in Steel Production Systems

Abstract

A wire rod and bar rolling process is important in steel production systems. Its scheduling problem involves the constraints on sequence-dependent family setup time and release time. This work intends to schedule the batches with multiple jobs to minimize the number of late jobs. An important characteristic of this problem is that the number of late jobs within a batch varies with its start time. Given a start time of a batch, the number of late jobs within it can be derived. Two problem-specific scatter search algorithms are developed to solve this problem. They are tested on a benchmark with 120 instances whose optimal solutions are known and compared with an exact method. It shows that they can effectively solve the concerned problem and are much faster than the exact method for larger-scale cases. Their usage can well meet the industrial scheduling needs arising from a wire rod and bar rolling process while an exact method fails.