Dear pgRouting community,
I am Moaaz Tarek, a computer engineering senior student at Cairo University.
Lately i have been interested in route optimization especially the DPDPTW ( dynamic pickup and delivery problem with time window) , also i am eager to contribute to vrpRouting as a GSOC participant.
i have been researching for 2 month about the PDP and its variants and i have come across different solutions methods that i would like to implement
- MOEA/D: a multi-objective evolutionary algorithm with diffusion that sets multiple objectives and try to find a feasible solution using VNS as its local search method to avoid local min/max
- Ant colony system with bipartite graph matching: this variant solves the DPDPTW with dynamic customer arrival. It introduce a bipartite graph matching model within the neighborhood search process and employ the KuhnMunkres (KM) matching algorithm, generally it does a Large Neighborhood search as it local search method
Research papers list that these methods provide better solution in less time than the tabu search
i hope this idea would be a good fit to vrpRouting, and I would like to hear your opinions.
Thank you,
Moaaz