An improved Partheno-genetic algorithm for travelling salesman problem

This paper presents an improved Partheno-genetic algorithm (IPGA) for the travelling salesman problem (TSP). IPGA repeals the crossover operators used generally by traditional genetic algorithms, where the genetic operation only happens in one chromosome. IPGA can reduce the number of individuals by improving the method of describing individuals and can also decrease the computing time by improving the method of computing individuals' fitness. For IPGA, the individuals among original populations are handled so that the average fitness of the individuals is increased. A variety of individuals among the original populations can be maintained by checking the difference of individuals and removing part of the same individuals in the course of genetic operation. The proposed measures can increase the convergence rate and improve the ability to search in a global space. Simulating examples show the effectiveness of the IPGA method.