An Accessibility-Oriented Optimal Control Method for Land-Use Development

Abstract

The performance of transport systems is seldom considered when designing land-use development planning, which could result in a deteriorated transport system. One of the major goals of a transport system is to improve individuals’ accessibility or the ease of reaching desired activities, destinations, and services. For urban smart growth, policymakers usually face a challenge of how to coordinate land-use development in multiple urban centers in order to optimize accessibility. This paper aims to provide an optimal land-use development method in terms of system accessibility. A novel bilevel model system is proposed to represent the problem where the upper level aims to optimize transport system performance in terms of accessibility, and the lower level is a four-step model with feedback to achieve transport system equilibrium. A multinomial logit model is used for destination choice, which can adequately capture traveler decision behaviors compared with the gravity-analogous model. To find the optimal solution of the proposed bilevel model, an efficient algorithm is proposed on the basis of a Dirichlet distribution, method of successive averages (MSA), Frank-Wolfe algorithm, and Dijkstra algorithm. An experimental study using the Nguyen-Dupuis network is presented to verify the effectiveness of the proposed model and solution algorithm. The research demonstrates that the optimal control of land-use development is critical to network accessibility, and the modeling framework can be a useful tool in determining urban development policies.