This paper is aimed at the selection of optimal configurations for power systems on oil platforms for generating electricity and satisfying power demands of process equipment. The variable power demands arising from the variable production profile are approximated through a finite sequence of time periods. A systematic procedure for the design of these flexible power systems is presented that is based on a multiperiod mixed-integer linear programming model. The model, which has as an objective to minimize the net present value of the investment and operating cost, can select among several types of gas turbines for satisfying the power demands, and it also considers the optimal assignment of drivers for rotary process equipment. The application of this model is illustrated on the power system on an oil platform in the North Sea. The problem consists in selecting gas turbine type as driver for the generator, and to decide what kind of driver, gas turbines or motor, is going to be used at the two selected equipment nodes. The optimal selection was obtained in approx 234 sec of CPU time on a OEC-20 computer. The type choosen is the Solar Mars gas turbine, which was the smallest of the gas turbines. The two nodes are both driven by electrical motors.