This paper presents an evolution program for optimal expansion problem with backlogs in demand which is to determine when and how much capacity should be added under varying types of demand and cost. To overcome premature convergence and stalling, we employ an exponential-fitness scaling scheme. To improve the chromosomes, we introduce hetero-dimensional mutation which generates a new dimension and produces a feasible solution, and homo-dimensional mutation which mutates the chromosomes in the negative of gradient direction. Finally, a numerical example is discussed.