Effective and efficient scheduling of dynamic job shops—Combining the shifting bottleneck procedure with variable neighbourhood search

Minimizing total weighted tardiness for job shops is a hard production scheduling problem. Shifting bottleneck (SB) procedures, which use a backtracking technique, can achieve excellent results for this problem. However, they usually require long computation times and do not scale well to larger problem instances of a realistic industrial size. In order to enhance the efficiency of SB procedures, we use variable neighbourhood search (VNS) instead of backtracking to optimize the machine planning sequence and the decomposed single machine problems. Evaluation on static benchmark problem instances and more realistic, dynamic scenarios confirms the performance and efficiency of our SB-VNS procedure.