2005 Volume 56 Issue 5 Pages 366-376
Investment planning is the selection of low-cost generation plants to meet customer demands over a multi-decade time horizon. In uncertain business environments, it is important to make management decisions considering not only the optimality, but also the uncertainties in productive activity. Here, we focus on the electric powerplants, treat the demand and energy price in the future as uncertain elements, and show a robust model of electric powerplant planning. The proposed robust model is formulated as a 0-1 integer programming problem which maximizes operating profit while satisfying future demand and dealing effectively with the anticipated variation in operating profit. The 0-1 integer problem tends to become large-size for the electric powerplant planning of existing Japanese companies, and it is difficult to solve the large-size 0-1 integer problem with standard methods such as the branch-and-bound method. Thus, we propose an alternative solution method that splits the 0-1 integer problem into many linear programs using the dynamic programming technique and solves them sequentially. The numerical results show that small-scale powerplants are suitable to cope with severe uncertainty even if the operating costs per unit of production of the small-scale plants are higher than those of large-scale plants. The effectiveness of our method is demonstrated with various kinds of test problems which vary the number of electric powerplants and the period of the electric powerplant planning.
