New integer linear programming formulation for the traveling salesman problem with time windows: minimizing tour duration with waiting times


KARA İ., Koc O. N. , ALTIPARMAK BAYKOÇ F. , DENGİZ B.

OPTIMIZATION, vol.62, no.10, pp.1309-1319, 2013 (Journal Indexed in SCI) identifier identifier

  • Publication Type: Article / Article
  • Volume: 62 Issue: 10
  • Publication Date: 2013
  • Doi Number: 10.1080/02331934.2013.824445
  • Title of Journal : OPTIMIZATION
  • Page Numbers: pp.1309-1319

Abstract

The travelling salesman problem, being one of the most attractive and well-studied combinatorial optimization problems, has many variants, one of which is called travelling salesman problem with Time Windows (TSPTW)'. In this problem, each city (nodes, customers) must be visited within a time window defined by the earliest and the latest time. In TSPTW, the traveller has to wait at a city if he/she arrives early; thus waiting times directly affect the duration of a tour. It would be useful to develop a new model solvable by any optimizer directly. In this paper, we propose a new integer linear programming formulation having O(n(2)) binary variables and O(n(2)) constraints, where (n) equals the number of nodes of the underlying graph. The objective function is stated to minimize the total travel time plus the total waiting time. A computational comparison is made on a suite of test problems with 20 and 40 nodes. The performances of the proposed and existing formulations are analysed with respect to linear programming relaxations and the CPU times. The new formulation considerably outperforms the existing one with respect to both the performance criteria. Adaptation of our formulation to the multi-traveller case and some additional restrictions for special situations are illustrated.