Expected utility theory for the optimization of bridge management