This paper presents a new method for identifying optimal transport strategy, and compares it with some other approaches. Investigations were conducted, using a hypothetical strategic land-use transport planning model (PLUTO). They showed that the relationships between policy measures and key indicators of system performance, which these models predict, can be approximated by relatively simple regression models. The coefficients of these models can be used to help identify the best performing combinations of policy measures. The paper presents the most efficient method of specifying and calibrating these models. PLUTO is a menu-based interactive model of a hypothetical city, represents many types of mechanisms and traveller decisions, and can introduce a wide range of policies into the city. A numerical example for discrete projects is presented in some detail, comparing the relative merits of the following approaches: (1) a systematic approach; (2) an intuitive approach using the GLIM statistical modelling software; and (3) an intuitive approach not involving GLIM. For the continuous project case, the example is presented for a logical intuitive GLIM approach; comparison with a non-GLIM approach is not easy here. Some further work is indicated.
Abstract