Authors:
José Cáceres-Cruz
1
;
Daniel Riera
1
;
Roman Buil
2
;
Angel A. Juan
1
and
Rosa Herrero
2
Affiliations:
1
Open University of Catalonia, Spain
;
2
Universitat Autònoma de Barcelona, Spain
Keyword(s):
Heterogeneous Vehicle Routing Problem, Asymmetric Cost Matrix, Clarke and Wright, Randomized Algorithms, Heuristics.
Related
Ontology
Subjects/Areas/Topics:
Applications
;
e-Business
;
Enterprise Information Systems
;
Logistics
;
Operational Research
;
OR in Transportation
;
Pattern Recognition
;
Routing
;
Software Engineering
Abstract:
Urban transportation is a strategic domain that has become an important issue for client satisfaction in distribution companies. In academic literature, this problem is categorized as a Vehicle Routing Problem, a popular research stream that has undergone significant theoretical advances but has remained far from practice implementations. Most Vehicle Routing Problems usually assume homogenous fleets, that is, all vehicles are considered of the same type and size. In reality, this is usually not the case as most companies use different types of trucks to distribute their products. Also, researchers consider symmetric distances between customers. However, in intra-urban distribution it is more appropriate to consider asymmetric costs. In this study, we address the Heterogeneous Fixed Fleet Vehicle Routing Problem with some additional constraints: (a) Asymmetric Cost matrix, (b) Service Times and (c) Routes Length restrictions. Our objective function is to reduce the total routing cost
s. We present an approach using a multi-start algorithm that combines a randomized Clarke & Wright’s Savings heuristic and a local search procedure. We execute our algorithm with data from a company that distributes food to more than 50 customers in Barcelona. The results reveal promising improvements when compared to an approximation of the company’s route planning.
(More)