One commodity pickup and delivery traveling salesman problem and an extension formulations and algorithms
Abstract
In the One Commodity Pickup and Delivery Traveling Salesman Problem (1-PDTSP), we are given a set of nodes (cities) along with distance (cost) associated with each pair of the nodes. A special node is defined as the depot of the capacitated vehicle while other nodes are referred as customer nodes which require certain service (pickup or delivery). The goal is to minimize the total trip cost of the vehicle that departs from the depot, visits every node in the network exactly once, provides the service (pickup/delivery) at the visited node, and returns to the depot. The 1-PDTSP is a generalization of the Traveling Salesman Problem (TSP). There are many applications in freight transportation and logistics where these problems commonly arise. In this thesis, we explore two important problems related to the 1-PDTSP: (1) A basic version of the pickup and delivery problem which satisfies the demand and capacity constraints and (2) A variant of the basic version which incorporates the time window constraints. For each of these problems, we develop a mathematical programming formulation and discuss about the feasibility and optimality of the solution. Computation experiments are conducted to demonstrate the complexity of the optimal formulation. In addition, we develop a heuristic algorithm for each of the problems with the goal of addressing the computation complexity. The heuristic algorithms incorporate few well adopted techniques in the literature. Extensive computational results are presented for the heuristic approach and these are compared with the results from exact algorithms. Findings and recommendations along with future directions are also discussed in this thesis.
Degree
M.S.C.E.
Advisors
Ukkusuri, Purdue University.
Subject Area
Civil engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.