Using two reinforcement learning ideas to solve certain max-cut problems.
Max-cut problem requires us to find two disjoint and seperate point sets and at the same time to maximize weight sum of all cut edges.
In this small attempt, I focus on one certain undirected graph. If we want to generalize the solution, just change the input whose form is (v1,v2).
By default, the weight of all the edges are seemed as 1.
MCMC and MCTS are used to solve this problem.