Distributed optimal planning: an approach by weighted automata calculus