Skip to Main Content
A recent integration showed that ant colony optimization (ACO) algorithms with immigrants schemes perform well on different variations of the dynamic travelling salesman problem. In this paper, we address ACO for the dynamic vehicle routing problem (DVRP) with traffic factor where the changes occur in a cyclic pattern. In other words, previous environments will re-appear in the future. Memory-based immigrants are used with ACO in order to collect the best solutions from the environments and use them to generate diversity and transfer knowledge when a dynamic change occurs. The results show that the proposed algorithm, with an appropriate size of memory and immigrant replacement rate, outperforms other peer ACO algorithms on different DVRP test cases.
Evolutionary Computation (CEC), 2012 IEEE Congress on
Date of Conference: 10-15 June 2012