A fast interactive solution method for large capital expenditure selection problems

In many capital expenditure planning problems, a large number of alternatives compete for the expenditure of the capital budget. Standard integer programming approaches select alternatives to maximize total return, subject to financial and other constraints. For large problems, obtaining an optimal solution may be so time consuming, the slow turn-around time prevents extensive sensitivity analyses. Typically, the constraints in the problem are actually ‘soft’ and expending a large amount of computing effort to find an ‘optimal’ solution satisfying these constraints is not necessary. It is more important to understand the value of the constrained resources. It is argued here that, in many cases, the integer programming formulation is not appropriate. In the integer programming formulation, alternatives are selected to fill out the resource constraints and maximize total return, and some of these alternatives are generally not those that give the best return on resource use. By recognizing this, a solution method is developed that provides more useful solutions quickly. This allows for the development of an interactive decision analysis tool.