Commute-Related Activity (Trip) Chaining Analysis

The analysis of commute trip patterns has always been of considerable interest in the travel demand literature since the commute pattern has a decisive impact on peak period traffic congestion on roadways. Several studies have examined commute patterns from metropolitan areas in the United States in the past decade. A consistent finding of these studies has been that the commute pattern is becoming more complex due to an increasing tendency to make non-work stops during the commute, especially in the evening. Thus, it is important to understand stop-making behavior during the work commute. The broad objective of this research effort is to model the entire activity-travel pattern of the worker between the time s/he leaves work to the time s/he returns home at the end of the evening commute. The attributes characterizing the evening commute activity-travel pattern include: a) number of stops (including zero stops, which implies that the worker heads home directly), b) sequence of stops (if the number of stops is more than one) c) activity type of each stop, d) activity duration of each stop, e) travel time deviation to each stop from previous stop (or from work if the stop is the first one) relative to the direct travel time from previous stop to home (we will refer to this simply as the travel time deviation from previous stop), e) departure time from work, f) travel mode used for the commute, and g) location of each stop. Techniques being used in the effort include discrete-continuous models, hazard-based duration models, mixed-logit methods, joint unordered/ordered discrete choice models, and longitudinal/spatial methods. The model system is being used to examine the complex behavioral responses of commuters to transportation and land-use policy actions, and to evaluate the traffic and air quality implications of such actions.