Pedestrian Activity Assignment Problem with Time-Space Constraint and Path Correlation

Abstract

Spatial planning for pedestrians attracts attentions, and the methods to evaluate the planning based on pedestrian behavior are needed. However, pedestrian route choice behavior has difficulties for modeling; 1) a path enumeration problem because of the route variety, 2) pre-trip decision mechanism and trip-based model are not suitable since there is high continuity of walking and staying behavior in pedestrian contexts. In this paper, we propose a pedestrian activity path choice model with time-space constraint in time-structured networks. We formulate an implicit and network-GEV based model to describe activity correlation. The model allows us to compute the activity assignment with cyclic, multi-trip and time attributes. The assignment results show that time-space discount rate of expected utility and the prism constraint are important parameters which have an influence on pedestrian decision making.