Cogeneration systems (CGSs) consist of power generators and thermally activated machines used to recover the thermal energy from the generators. The CGS optimal design problem is an optimal location problem of facilities that produce energy through two stages. Electric power and thermal energy are produced in the first stage and thermal energy is converted to effective energy in the second stage. An allocation of facilities is evaluated as the sum of the initial cost associated with the equipping of facilities and the minimum running cost entailed in satisfying the energy demand. Therefore, the CGS optimal design problem is a difficult problem that is composed of an optimal location problem and an optimal scheduling problem. In this paper, we present a mixed integer programming (MIP) formulation of the CGS optimal design problem. The proposed formulation enables us to use a general purpose MIP solver that is readily available. We show that realistic CGS optimal design problems arising in a hospital, a hotel and an office can be solved with reasonable computational costs.