The genetic algorithms (GAs) are evolutionary algorithms, inspired by the biology, used often to solve problems classified as NP-hard combinatorial optimization problems, like the traveling salesman problem (TSP) which the challenge is to find a Hamiltonian cycle linking 𝑜 points while minimizing the target. The GAs do not provide an exact solution but they allow finding a solution very close to the optimum in a reasonable time. In this work, we will try to adapt the traditional principle of genetic algorithms (selection, crossover, mutation ...) and apply it to the mentioned problem in order to find good solutions for different benchmark instances from the TSPLIB library. To further improve the results obtained, we will introduce a new approach inspired also by nature, which is that of immigration. It is an approach which ensures a great diversity in the population and which will, therefore, obtaining satisfactory results.
You can find the full report,statistics and results obtained here 👉 Final Report.