In this paper a three-stage hybrid approach to generalized genetic algorithm is proposed that relies on a combination of Tournament and Best Solution methods for selection, PMX and Ordered crossovers and Inverse and Swap mutation to obtain optimal distance measures for the Traveling Salesman Problem. The three hybrid combinations are made in order to retain diversity of the population while maintaining elitism without leading to late convergence. The hybrid is then tested against TSPLIB. Experimental results showed that the approach leads to convergence rates as high as 100% for less than 100 cities and above 93% for a larger number of cities.