A combinatorial optimization approach is proposed to design systems that satisfy time-varying utility demands. A numerical example illustrates the application of the procedure and is used to compare the traditional simulated annealing algorithm with a nonequilibrium version.