Lai Soon Lee, Kien Hua Ting, Hsin-Vonn Seow

    This paper is motivated by the unbalance utilization rate of public transit which affects the take-up rate of public transportation. The authors advocate solving this via the first-mile ridesharing problem. The selective open vehicle routing problem is used to model the first-mile ridesharing problem. Constrained K-Mean and K-Mean clusterings are used to cluster the dataset to represent the number of available drivers to service the transit to the stations. In terms of the meeting point selection, it can either be at a mutual meeting point (centroid) or at one of the cluster’s commuter residences (non-centroid). For this, two types of models, Multi Origin Single Destination Split Delivery Open Vehicle Routing Problem (MOSD-SDOVRP)(Centroid) and Multi Origin Single Destination Split Delivery Selective Open Vehicle Routing Problem (MOSD-SDSOVRP)(Non-Centroid), are discussed. The proposed models are evaluated and compared using CPLEX with the well-known Solomon benchmark dataset. The results will allow a smooth transit for commuters from their respective residences to the station to encourage a high take-up of public transportation.

Key words:public transit;ridesharing;first-mileproblem;split delivery.

DOI:10.11113/matematika.v45.n1.1229

Date:2023-6-27