Ant Colony Approach for Multiple Pickup and Multiple Dropoff

Abstract

The Multiple Travelling Salesman Problem, popularly known as MTSP is an NP-hard problem. MTSP is a well-known combinatorial optimization problem in which more than one salesmen visit all cities only once and return to the depot. In our problem, we apply the MTSP algorithm to multiple drivers picking and dropping packets at multiple locations and the drivers not returning to the starting location. There are no exact solutions for solving this combinatorial problem that can guarantee to find the optimal route within a reasonable time. A meta-heuristic algorithm, Ant Colony Optimization (ACO) is used as a base for our solution construction for different variations of the problem such as handling multiple pickups and multiple drop-offs using a single driver, multiple drivers, drivers starting at different times, and drivers available for different times. The goal is to maximize the number of goods delivered while minimizing distance (or time) within some threshold limits. The results are compared to existing algorithms like Brute-force approach and Nearest Neighbor algorithms. Our results show that the proposed ant colony algorithm achieves better results or at worst identical results to the Brute-force approach.