Ant colony system for carpool service problem with high seating capacity

Abstract

Carpool Service Problem (CSP), which aims at providing the ridesharing plans in order to alleviate traffic congestion, has attracted high attention in the past years. A considerable amount of efforts have been devoted to solving CSP with low vehicle capacity and small travel demand, e.g., ridesharing by private cars. However, there are few studies involving CSP with high vehicle capacity, e.g., ridesharing by buses or microbuses. In view of the special high capacity characteristic, this paper proposes an ant colony system (ACS) with novel heuristic information and pheromone calculation strategies (HC-ACS). First, we redesign the heuristic information of each edge, both the length of the specific edge and the estimation of the total travel are considered. Second, a summation rule is applied to the usage of pheromone to maintain the searching diversity. Our experiments on datasets with different spatial distribution show that the proposed HC-ACS is promising in the environment with relatively high seating capacity.