This paper develops and tests a mixed-integer programming model for finding the optimal timing and sizing of plant capacities, The object of the model is to minimize the total discounted costs over a planning horizon. The factors that are considered for this purpose are the current and future values of the following: discount rates, construction and setup costs, operating costs, subcontracting costs, idle costs of excess capacities, and growth rates of demand. Actual data from a set of electrical power generating facilities are utilized to test the model. Several variations in time paths of demand, and the cost factors are considered to assess the effects of each factor on the timing and sizing decisions