Designing multiple short-turn routes to mitigate the crowding on a bus network

Abstract

On a transit route, initiating a trip beyond a transit route’s departure terminus and terminating the trip before its arrival is called ‘‘short-turn.’’ This study develops a method to design short-turn services to alleviate crowding on a transit network consisting of multiple overlapping routes. The proposed method consists of two processes: (i) optimizing routes, turn-back points, and fleet sizes of short-turn services by a genetic algorithm to minimize the sum of waiting, in-vehicle, operation, and social costs; (ii) assigning passenger trip demands to short-turn and existing services. These processes are performed iteratively to design multiple short-turn routes. The proposed method is applied to a real-world transit network in Seoul, Republic of Korea, to design 10 short-turn routes. The results are compared with two naïve methods that introduce shortturn services to sections with the maximum occupancy and the maximum demand divided by the section length, respectively. The comparison results show the proposed method is much superior to both of the naïve methods in reducing total costs. The result of the sensitivity analysis suggests that the value of in-vehicle time needs to be estimated accurately and site-specifically. The proposed model contributes to enhancing the convenience of transit users by effectively mitigating the crowding under budget constraints.