You can follow this article http://ieeexplore.ieee.org/xpl/abstractMetrics.jsp?arnumber=1259748&tag=1

Please check my paper at : Proceedings of the 4th Indian International Conference on Artificial Intelligence, IICAI 2009, Tumkur, Karnataka, India, December 16-18, 2009; 01/2009. You can download paper from my contributions page. Hope this will help you.

Honestly, why would you want to use a technique designed for continuous domains on discrete, combinatorial problem?

I always hear that PSO has been proposed with continuous optimization in mind and that because of that it is suboptimal for combinatorial optimization. I don't agree, or else evolutionary algorithms in general should never be applied to combinatorial optimization. I do agree, however, that the adaptation has to be done carefully.

I recommend you check the different papers referred to here, and I'll add another one that was the first one I read about PSO: 

Chapter Particle Swarm Optimization Algorithm for the Traveling Sale...

https://www.researchgate.net/publication/297245624_Particle_Swarm_Optimization_Algorithm_and_its_Codes_in_MATLAB

Research Particle Swarm Optimization: Algorithm and its Codes in MATLAB

https://www.researchgate.net/publication/296636431_Codes_in_MATLAB_for_Particle_Swarm_Optimization

Code Codes in MATLAB for Particle Swarm Optimization

https://www.researchgate.net/publication/280244004_A_comparative_study_of_metaheuristic_optimization_approaches_for_directional_overcurrent_relays_coordination

Article A comparative study of metaheuristic optimization approaches...

It is not recommended to do in this way. But if you are interested, you can reperesent every indivuidual by a complete tour of graph. Clearly, it will be hard and time consuming to implement PSO search operations on this type of individuals.