Heuristic approaches for Scheduling jobs and Vehicles in a Cyclic Flexible Manufacturing System
Sprache des Vortragstitels:
Englisch
Original Tagungtitel:
EURO 2018 - 29TH European Conference on Operational Research
Sprache des Tagungstitel:
Englisch
Original Kurzfassung:
This paper considers the Scheduling of Automated Guided Vehicles (AGVs) in a flexible flow shop environment. The AGVs travel along a single loop. All machines are set alongside the track in the correct order, with multiple machines per stage. All AGVs are to be scheduled for a specific starting time and will then circle the track without stop. Pickup and delivery times are included in the travel time of a vehicle between two machines. Jobs may start as soon as they have arrived and their predecessor has been started for processing. Therefore, job completion times are dynamic.
The considered objectives are the minimization of the number of AGVs and the minimization of the makespan. For the resulting problem, different metaheuristic search procedures are proposed and compared among each other as well as a basic local search algorithm relying on small changes to the AGV starting times. Optimal results are produced by means of a brute force enumeration algorithm. Finally, fixed permutation schedules are compared to processing jobs according to the first-come-first-serve rule.
Sprache der Kurzfassung:
Englisch
Englischer Vortragstitel:
Heuristic approaches for Scheduling jobs and Vehicles in a Cyclic Flexible Manufacturing System
Englischer Tagungstitel:
EURO 2018 - 29TH European Conference on Operational Research
Englische Kurzfassung:
This paper considers the Scheduling of Automated Guided Vehicles (AGVs) in a flexible flow shop environment. The AGVs travel along a single loop. All machines are set alongside the track in the correct order, with multiple machines per stage. All AGVs are to be scheduled for a specific starting time and will then circle the track without stop. Pickup and delivery times are included in the travel time of a vehicle between two machines. Jobs may start as soon as they have arrived and their predecessor has been started for processing. Therefore, job completion times are dynamic.
The considered objectives are the minimization of the number of AGVs and the minimization of the makespan. For the resulting problem, different metaheuristic search procedures are proposed and compared among each other as well as a basic local search algorithm relying on small changes to the AGV starting times. Optimal results are produced by means of a brute force enumeration algorithm. Finally, fixed permutation schedules are compared to processing jobs according to the first-come-first-serve rule.