This paper presents some results of an on-going effort to simulate the process of the individual trip linking. The previous contributions in this field suggested the stochastic models based on the formal assumption that the future trajectory of an individual trip linkage depends only on its present state. The idea of the trip linkage as a memoryless behaviour is questionable, and there might be a high risk of error in its application for planning purposes. The research reported attempts to express explicitly and to check at disaggregated level, the interaction of all trips and stops in an individual's schedule. In general, this aims at accounting for the interrelationships of many trip-linkage characteristics, such as category, frequency, time, duration and distance. Within this framework, both the modelling and information on the reality are progressing gradually. As a specific part of the study, the model presented was started from the information on the individual frequency of non-home arrivals to simulate a distribution of all arrivals over the chains of different sizes. This model is based on the harmonic series concept and generates the output in various forms: production and distribution of chains for the given non-home sojourn demand; mean number of trips for n calls, as well the mean number of remaining trips after the i-th call out of n calls. The model has been tested with a one-percent sample from the population of south amsterdam. The first results of the test are encouraging. The findings of the study could be successively incorporated in the current algorithms of urban transportation study. In this stage, the main advantages of the approach suggested would be the accounting for the continuity of individual schedules, and the replacement of conventional trip generation and attraction measures by the non-home arrival production which might be a better measure for the real travel need. (a) for the covering abstract of the symposium, please see irrd abstract no. 224453.
Abstract