The implementation of Niching Memetic Algorithm for Evolutionary Diversity Optimization on Traveling Salesperson Problem, as described in the work:
Do, A.V., Guo, M., Neumann, A. and Neumann, F. (2022). Niching-based Evolutionary Diversity Optimization for the Traveling Salesperson Problem. Proceedings of the Genetic and Evolutionary Computation Conference. DOI: 10.1145/3512290.3528724
Within the 2022_GECCO folder contains MATLAB implementation of NMA for EDO, as well as simple mutation-based EA for EDO. The chosen representation is visit-order permutation.
- run.m
The entry point where the experiment is run, containing settings with evaluation budgets, threshold values, etc. - tsp_instances.mat
Contains data from 10 TSPLIB instances, including name, distance matrix, 2d Euclidean coordinates of vertices, and known optimal solution. - div_tsp_p1.m
The NMA as a function. - dived.m
(mu+1)-EA equalizing edge distances, maximizing sum-sum diversity. - divpd.m
(mu+1)-EA maximizing smallest pairwise distances, maximizing sum-min diversity.
Run run.m as is to replicate the results. The output should be written into a separate file.