This paper proposes a swarm intelligence approach based on a disjunctive graph model in order to schedule a manufacturing system with resource flexibility and separable setup times. Resource flexibility assigns each operation to one of the alternative resources (assigning sub-problem) and, consequently, arranges the operation in the right sequence of the assigned resource (sequencing sub-problem) in order to minimize the makespan. Resource flexibility is mandatory for rescheduling a manufacturing system after unforeseen events which modify resource availability. The proposed method considers parallel (related) machines and enforces in a single step both the assigning and sequencing sub-problems. A neighboring function on the disjunctive graph is enhanced by means of a reinforced relation-learning model of pheromone involving more effective machine-sequence constraints and a dynamic visibility function. It also considers the overlap between the jobs feeding and the machine (anticipatory) setup times. It involves separable sequence-independent and dependent setup phases. The algorithm performance is evaluated by modifying the well-known benchmark problems for JOB shop scheduling. Comparison with other systems and lower bounds of benchmark problems has been performed. Statistical tests highlight how the approach is very promising. The performance achieved when the system addresses the complete problem is quite close to that obtained in the case of the classical job-shop problem. This fact makes the system effective in coping with the exponential complexity especially for sequence dependent setup times.

Flexible Job Shop Scheduling with Sequence-dependent Setup and Transportation Times by Ant Colony with Reinforced Pheromone Relationships

ROSSI, ANDREA
Primo
2014-01-01

Abstract

This paper proposes a swarm intelligence approach based on a disjunctive graph model in order to schedule a manufacturing system with resource flexibility and separable setup times. Resource flexibility assigns each operation to one of the alternative resources (assigning sub-problem) and, consequently, arranges the operation in the right sequence of the assigned resource (sequencing sub-problem) in order to minimize the makespan. Resource flexibility is mandatory for rescheduling a manufacturing system after unforeseen events which modify resource availability. The proposed method considers parallel (related) machines and enforces in a single step both the assigning and sequencing sub-problems. A neighboring function on the disjunctive graph is enhanced by means of a reinforced relation-learning model of pheromone involving more effective machine-sequence constraints and a dynamic visibility function. It also considers the overlap between the jobs feeding and the machine (anticipatory) setup times. It involves separable sequence-independent and dependent setup phases. The algorithm performance is evaluated by modifying the well-known benchmark problems for JOB shop scheduling. Comparison with other systems and lower bounds of benchmark problems has been performed. Statistical tests highlight how the approach is very promising. The performance achieved when the system addresses the complete problem is quite close to that obtained in the case of the classical job-shop problem. This fact makes the system effective in coping with the exponential complexity especially for sequence dependent setup times.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11568/628464
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 80
  • ???jsp.display-item.citation.isi??? 67
social impact