Elise Miller-Hooks

Publications: Journal 19

“Solving a Generalized Traveling Salesperson Problem with Stochastic Customers”

Computers and Operations Research 34, 1963-1987.
Tang, H. and E. Miller-Hooks (2007)

Pub19In this paper, a probabilistic generalized traveling salesperson problem (PGTSP) is introduced to address several applications. In the PGTSP, each customer belongs to a cluster that consists of a set of customers. Whether or not any given customer will be present during actual operations is known a priori only probabilistically. The PGTSP seeks the minimum expected length tour to visit a subset of all customers such that the tour traverses each cluster at least once. If when implementing the tour it is revealed that there is no demand for service within the cluster to which the stop belongs, that stop will be skipped. An exact solution algorithm based on the integer L-shaped method and three tour construction-based heuristics for quickly solving this problem are described in the paper.

Computational experiments were conducted to assess computational requirements and solution quality of the proposed solution techniques. These experiments show that the exact method is able to solve small- and moderate-size problems to optimality. In addition, one of the proposed heuristics (the MMI heuristic), in particular, gives good approximate solutions (often a few percent from optimal) in very reasonable computational time.

Elise Miller-Hooks, Ph.D.
Bill & Eleanor Hazel Chair in Infrastructure Engineering

Phone: 703.993.1685
Email: miller@gmu.edu

Office: 4614 Nguyen Engineering Building

Sid and Reva Dewberry Department of Civil, Environmental and Infrastructure Engineering
George Mason University
4400 University Drive, MS 6C1
Fairfax, VA 22030


More Publications

Volgenau School of Engineering
George Mason University