A Transit Route Network Design Considering Equity

Abstract

Due to the polarization of incomes, low birth rate and aging population in modern society, there is a significant emphasis on social equity as imbalances deepen and conflicts become more frequent. Public transportation also has temporal and spatial imbalances, which are becoming more serious as faster and more direct service is provided between high-demand regions. Therefore, there is a need for an adjustment process. A recently developed new towns in suburban areas can be called as representative areas being affected by the operation strategy of public transportation with the purpose of minimizing costs. In this study, the public transport network assessment methodology considering modal and spatial equities is proposed. Here, equities are considered by two indexes which can reflect mobility and accessibility simultaneously. Also, heuristic transit route network determination process with sequential procedures which consist of target line selection, target node selection, and alternative line determination is proposed, and frequency setting procedure is implemented. The model is configured through the bi-level modeling based on an iterative process in order to calculate the modal split and traffic and transit assignment with transit route network changes. Based on the transit route network taking account of the level of modal and spatial equities, in the upper model, frequency of each line is determined by genetic algorithm. On the other hand, in the lower model, modal split and traffic and transit assignment are implemented. This transit route network design model and algorithms are applied on a sample network. As a result, improved solution with equity and total cost was found based on the comparison with existing transit route network. There are some expected contributions in this study. First, equity indexes are proposed to assess the mobility and accessibility. Second, iterative transit route network determination methodology to maximize equity is proposed. Last, genetic algorithm to determine the frequency of each line is improved with practical frequency constraint, modal split and assignment with variable demand are reflected.