hllqna
commented
2 weeks ago I'm glad that you've found my implementation inspirational for further improvements of your algorithm, good luck! I'd like to point out that in my programme, during the local search phase, there is no restriction on the possible states generated with a tweak function, so it doesn't follow the idea of the tabu search. However, it's also an interesting approach worth exploring
There has been significant attention to building the initial solution. The build_init_solution() function offers an interesting approach to tackling the "set cover" problem. I have also considered duplicates and sets that cover a single element. However, the idea of adding sets iteratively and checking the validity of the coverage could make my algorithm more robust.
I would have liked to implement Tabu Search to solve the set cover problem, but I had already taken another direction. Nonetheless, I am pleased to have seen the results achieved using this methodology.