W. Fan

Pages: 43-54

Abstract
As an innovative form of project financing in which a private entity receives a concession from the public sector to finance, design, construct, and operate a facility stated in the concession contract, build–operate–transfer (BOT) offers unprecedented benefits and have been gaining popularity worldwide. The purpose of this paper is to present the modeling of multiple BOT projects in network equilibrium. These projects are assumed to have been controlled by more than one private company. Thus, the effects of making combined and separate decisions of the companies on toll charge and capacity are investigated. In order to achieve this, two different genetic algorithms (GA) based bi-level optimization solution methodologies are developed to determine the optimal toll charge and capacity of the new BOT links. Numerical experimentations are conducted by considering two BOT projects, which are controlled by two companies, in the famous Sioux Fall example network. For simplicity, the demand is assumed to be fixed and given a priori and the network consists of homogenous road users whose value of times (VOT) are uniform. The two scenarios are compared with each other in terms of the total system travel time (cost). The study shows that making a combined decision on toll charge and capacity of the new BOT links gives a better network performance than making a separate decision.

Keywords: optimal pricing; highway network improvement; build-operate-transfer; genetic algorithms