Distributed Markov Chain Redesign for Multi-Agent Decision-Making Problems