Gravity Model Calibration in the Four-Step Model of Transport Planning

Juan Miguel C. Manalo

Adviser: Adrian Roy L. Valdez, PhD, John Justine S. Villar, MSc

Abstract

With the Metro Manila traffic problem causing billions of pesos in losses due to mismanagement and congestion, a gravity model calibrated to suit the region's road network is required.

Thus, the following question is raised: What would be the appropriate parameter values to use and the constraints to subject the models into, and how valid would the resulting calibration be for the rest of the Four-Step Model (trip generation, trip distribution, modal split, trip assignment)? These would result in a trip distribution representative of Metro Manila traffic. The focus will be on the second step, calibrating on the entropy maximization model (EMM) using deterrence functions suggested by Ortuzar and Willumsen.

Two original metrics will be used: efficacy (for the overall transport population) and efficiency (for all modes). Subject to the constraints of the EMM, it is found that the exponential form [f(c_{ij}) = e^{-beta c_{ij}}; beta = 0.5] of the deterrence function is the most efficient when used with the gravity model. It is further supported by the balanced trip distribution and its plausibility in the trip assignment step. Furthermore, the three modes used (car, bus, jeep) are representative of the Metro Manila traffic system without taking time as a factor.

Further refinement and research is recommended.