Bus Transit Network Structure Selection With Multiple Objectives

2. Multiobjective Analysis In Transit Network Planning And Design

An exhaustive review of the multiobjective transportation network design problems was given by Current and Marsh (1993) and Current and Min (1986). The objectives formulated or suggested in these reviews address transportation costs, construction costs, distance, travel time, inventory, safety, household relocation, demand satisfaction, accessibility, quality, flexibility and reliability of service, facility utilization, profit, economic development and environmental concerns among others. The complexity of the objectives and the difficulty or impossibility of measuring them in commensurate units are also evident from these reviews. In general, multiobjective analysis is designed for problems in which the objectives cannot be readily measured in commensurate units. Moreover, multiobjective programming and planning is concerned with decision making problems on which there are several conflicting objectives (Cohon1978).

The model given by Rea (1972) is possibly the first of its kind that gives an assignment type of model for passenger flow concentration. This model takes advantage of the special characteristics of public transportation systems. It allows the planner to explore easily the combination of modes as a tool for generating and planning public transportation networks. The aim of the flow concentration model is to leave out some of the links with low flows. A similar approach was suggested by Hasselström (1981). Marwah et al. (1984) proposed a method that minimizes the total operation and riding time cost of the network. Carraresi et al. (1996), proposed a mathematical programming model for the regional mass transit (bus and rail) assignment. This model aims at improving the effectiveness of the mass transit system by modifying the departure trips taking into account the feasibility of the vehicle and driver schedules. Blue et al. (1997) presented a bi-objective path search algorithm. The objectives considered by this algorithm are travel time and trip complexity. Subsequently a tradeoff parameter is used to generate non-dominated solutions that can help in in-vehicle route guidance systems (IVRGS).