Figure 4 goal is to level the total load of all users instead of just one type of user. The company and each type of user would have a utility function reflecting its overall profit/cost. The utility functions are listed below, where wu, is for the utility company; ur, u2p, and war are for residential, commercial, and industrial users, respectively: ind f(Z) is the cost due to fluctuation of total user loads. spx, 3Br, and sp, are satisfaction functions for residential, com- nercial and industrial users, respectively. These functions have lifferent parameters based on the characteristics of the users, uid may take different forms other than (4) if necessary. In his paper, we employ the same satisfaction function as in the single-user-type case with different parameters, and the users’ yptimal responses to prices are in a form similar to (16), with lifferent parameters € and 7. The optimal responses of different vnes of users can be obtained as follows: