Robustness optimisation of the minimum makespan schedules in a job shop

In general, many minimum makespan schedules exist in job shops. Therefore, it is important to select the best one based on secondary criterion. In this paper, we adopt a measure called robustness of the schedule as the selection criterion. The robustness of the schedule is the maximum value of the sensitivity, which is a measure for the delay of completion times by one time unit, for all of the operations. As there are generally many minimum makespan schedules, it is difficult to enumerate all makespan-minimum active schedules and then to find the most robust schedule. A branch and bound method is introduced to the minimisation of makespan while optimising robustness based on a disjunctive graph model and the propositions proposed by Carlier and Pinson. A lower bound of robustness for each partial schedule is calculated using the corresponding disjunctive graph. The effectiveness of the proposed approach is clarified by solving test problems.