Microsimulation is increasingly assuming a major role in the advancement of demand-modeling practice. At the same time, it is attracting growing attention from the larger transportation-planning community. Four basic advantages of microsimulation versus conventional fractional-probability models are examined. The first is the technical advantage related to computational savings in the calculation and storage of large multidimensional probability arrays. The second is the meaningful advantage gained in the explicit modeling of various decision-making chains and time-space constraints on individual travel that allows for behavioral realism in the demand-modeling procedure. The third relates to the variability of microsimulation outcomes, which can yield full information about the distributions of the travel demand statistics of interest rather than single deterministic estimates or average values. As soon as constraints are introduced into the modeling framework (which often is done at the destination choice stage), competition arises, although generally it has been ignored in standard models. Microsimulation has the potential to handle this competition over work attractions and other travel activities in a meaningful fashion, which is the fourth advantage. These four advantages of microsimulation are discussed in light of the recent development and application of the New York best practice model, a microsimulation demand-modeling system for the New York-New Jersey-Connecticut metropolitan area.
Abstract