A decomposition algorithm to solve dynamic city-scale ridesharing problem

Abstract

This paper proposes a computational efficient, rapid, dynamic ridesharing algorithm. Firstly, a designed pre-processing procedure reduces size of the input date of potential nodes for vehicle to pass through. The proposed decomposition algorithm decomposes the large-scale problem with pre-processing date into independent sub-problems, and the method of parallel computing is adopted to solve the sub-problems. We call the dynamic programming algorithm to deal with the vehicle routing problem involved in sub-problems. Greedy random adaptive search algorithm is compared with the decomposition algorithm to verify the algorithm. Finally, a mathematical example shows that the algorithm can effectively solve large-scale dynamic ridesharing problems.