Model Information eXchange System (MIXS): A Unified Network Database Framework for Travel Demand Modeling

摘要

Many travel demand forecast models operate at state, regional and local levels. While they sharethe same physical network in overlapping geographic areas, they use different and uncoordinatedmodeling networks. This creates difficulties for models to exchange common information andcan result in data collection redundancies and difficulties in finding data errors or comparingfuture projections. This research investigated the issue of network information exchange amongmodels and proposed a framework to accomplish the information exchange using a unifiedstatewide GIS network approach. Named ‘Model Information eXchange System’ (MIXS), theproposed solution includes a geospatial relational data model that can support all participatingmodels with their input variables and forecast scenarios as well as protocols and operations toguide the exchange process. Two tests, one with a small database and another with a fullstatewide model, confirmed the feasibility of the MIXS database and processes. Althoughparticipation in MIXS is not without challenges, most technical problems considered can besolved. Successful implementation of MIXS will require a one-time time conversion to theunified network for any model to participate in MIXS. MIXS will create an environment thatpromotes convergence, standardization and unification of data and potentially modelassumptions, reduction of duplicate data collection efforts, reduction of errors, and ultimatelywill result in better and more consistent models throughout the state. A potential linkage of MIXSwith a cloud-based modeling engine is recommended as one of the future items to explore duringits practical implementation.