Date of Award
Doctor of Philosophy
This thesis concerns two routing problems, the 'Multiobjective Vending Problem' (abbreviated to MVP) and the 'Time Dependent Vending Problem' (abbreviated to TDVP).;To date, most research that deals with the problem of routing to cover a set of demand nodes has utilised a single objective approach, the objective being usually that of minimising distance or travel time. The set of nodes to be visited has been assumed known and specified. The MVP problem definition, based on a multiobjective solution approach, drops the latter assumption. The overall objective becomes that of identifying the trade-off relationship between two objectives, one to minimise some expression of route length, the other to maximise the coverage of nodes.;The study commences by discussing the advantages to utilising a multiobjective approach to optimisation research, stressing its potential role in spatial analysis. A number of general multiobjective research techniques are introduced. The MVP problem is defined mathematically, and a number of different solution approaches are discussed. Given present computing capabilities, solution by a heuristic based on the 'Constraint Method' is singled out as the most feasible approach to solve large MVP problems. Such a heuristic is designed, and is evaluated on a 25 node problem.;Problems of routing to cover a set of demand points have to date also predominantly focussed on problems where demand is uniform through time. The TDVP problem definition drops this assumption, allowing demand potential at the different nodes to vary with time. Times of arrival at the nodes become an explicit consideration in the problem formulation, the objective being that of identifying the optimal route that maximises demand potential covered. A second heuristic is designed to solve this problem, and is again evaluated on a 25 node problem.
Keller, Carl Peter, "Multiobjective Routing Through Space And Time: The Mvp And Tdvp Problems" (1985). Digitized Theses. 1452.