Differential evolution (DE) is a global optimizer for continuous design variables. To enhance DE, it is necessary to handle discrete design variables. In this paper, a discrete differential evolution (DDE) algorithm is proposed to handle discrete design variables. The proposed DDE is based on the DE/1/rand/bin method. In the proposed DDE, the mutation ratio is regarded as the exchange probability, and thus, no modifications of DE/1/rand/bin are required. In addition, in order to maintain diversity through the search process, we initialize all search points. By introducing the initialization of all search points, global or quasi-optimum solution can be found. We validate the proposed DDE by applying it to several benchmark problems.