I have a universe of N assets from which I need to optimally find the weights of (pre-set) K assets which produce the best Sharpe Ratio (fitness function).
Any links/references to research papers with code would be greatly appreciated?
I would first say that you will not (solve the problem using a meta-heuristic). There might be better alternatives to PSO. For examples, see
https://medium.com/@trbd/cardinality-constrained-portfolio-optimization-3e63e02454de
Preprint A Relaxed Optimization Approach for Cardinality-Constrained ...
http://liu.diva-portal.org/smash/record.jsf?pid=diva2%3A814624&dswid=-5601
If I were you I would not start by implementing a heuristic, when there are dedicated solvers/methods for the problem.
Please perform numerical tests so that you can verify whether mathematical programming approaches or metaheuristics perform better.
If you don't you will lose a potentially great improvement, and isn't an improvement what you would like???