Interactive, mixed-integer goal programming application of capital budgeting at Cities Service Company

Abstract

Scope and Method of Study: This study presents the development and implementation of a mixed-integer goal programming model to assist Cities Service management in strategic capital budgeting. Two types of input to the model were:

1. Corporate information and parameters - provided by corporate management; and

2. Strategic Planning Unit scenario information - provided by the unique corporate segments within Cities Service.

Given these inputs the desired output was an analysis of corporate direction via annual statistics in areas such as new debt issued and investments using the optimal scenario per strategic planning unit to maximize various corporate goals. A branch and bound algorithm was developed to assist in choosing the single scenario per strategic planning unit. All equations were linear in the model.

Findings and Conclusions: Six programs were developed and linked together in an interactive environment to produce corporate statistics over a ten year horizon. The third of these programs was an existing goal program. This program was combined with a branch and bound program, also developed, to produce the required mixed-integer solutions. Another output of the system was a report analyzing goal achievement. Four goals were allowed per run for which weights, priorities, and actual target values must be entered. The areas allowed as goals were net income, growth, return on assets, and assets. As described, the inputs and outputs are especially tailored to meet Cities Service Company's specific capital budgeting information requirements. Though conversion on the optimal solution was not attained as quickly as desired, various running options provided management with ample flexibility and convenience to meet their specifications.