System for Logical Verification and Inference of Activity (SYLVIA) Diaries

Abstract

The calibration of activity-based models requires the collection of activity diaries. Although activity diaries offer some potential advantages over conventional travel surveys, they involve a substantial burden on respondents, who are prone to making mistakes. Consequently, activity diaries tend to contain incomplete information and many inconsistencies, which is a serious problem because activity-based models require consistent and complete diary information. However, many logical relations can be imposed on the various choice dimensions underlying activity diaries, and further logical constraints and relations can be derived by combining activity diaries with data about the physical and transportation environment and the institutional context. Several rules may then be used to verify the consistency of observed activity diaries and to induce complete activity diaries in the case of missing information. Since the choice dimensions underlying activity diaries are highly interrelated, the links between the rules may become rather complex, implying that algorithms are required to induce consistent and complete activity diaries. Moreover, in many situations, alternative rules may be imposed. To support this task, an interactive computer system for the logical verification and inference of activity (SYLVIA), written in C++, has been developed. The system was primarily built to support the calibration of the ALBATROSS model, an activity-based model currently under development for the Dutch Ministry of Transport, Public Works and Water Management, but it is sufficiently generic to verify other activity diaries.