Consensus Q‐Learning for Multi‐agent Cooperative Planning