The Use of Constraints in Methods to Find Optimum Transport Policies
TIMMS P M, MAY A D, SHEPHERD S P, University of Leeds, UK
A substantial amount of research is now dedicated towards finding "optimal" sets of transport policies using a formal optimisation process; the paper by May et al (1998), describing the EU project FATIMA, gives a specific illustration of this method in ac
A substantial amount of research is now dedicated towards finding "optimal" sets of transport policies using a formal optimisation process; the paper by May et al (1998), describing the EU project FATIMA, gives a specific illustration of this method in action. At the heart of this method is a process whereby (real world) policy objectives are translated into objective functions, which encapsulate trade-offs between "good" and "bad" outcomes of any specific policy measure combination. This paper is concerned with "constraints" in these objective functions, which enable undesirable or infeasible sets of transport measures to be discarded in the optimisation process.
Section 2 gives an overview of types of constraints that can used, distinguishing between hard cons/ra#~/s and .~'@ co~s/rain/s. Comments are made of advantages and disadvantages of these two types of constraint, from both transport policy and practical computational viewpoints. Section 3 gives a summal 7 of the basic optimisation algorithm as used in the proiect OPTIMA which preceded FATIMA (Shepherd et al, 1997).
Section 4 lists three of the objective functions used in FATIMA, thus illustrating the importance of the concept of constraints. It also notes some of the computational defciencies of the OPTIMA optimisation process for the FATIMA objective functions, and suggests an improved algorithm which overcomes these deficiencies. Section 5 gives examples of the improved optimisation algorithms used in FATIMA, which attempt to overcome these deficiencies; whilst Section 6 draws some general conclusions.
Association for European Transport