A HIGH PRODUCTIVITY AND FLEXIBILITY MODELING METHODOLOGY FOR ENERGY AND ENVIRONMENT SYSTEMS

Recent remarkable progresses of computer technology and solving algorisms have enabled to solve large-scale optimization problems rather easily. Correspondingly, many large-scale models of energy and environment systems have been developed. However, it has become more difficult for modelers on one side to construct, modify and upgrade the models due to their complexity and large size, and for decision-makers on the other to understand the analyses results of the complicated models. Analysis of energy and environment systems usually requires large-scale models, and constructed models occasionally need to be modified or upgraded to meet changes in assumed data and in analysis purpose or to enjoy the progress of computer technology and of solving algorithms. Thus, a new methodology is wanted to provide a high productivity in model construction and a high flexibility to model modification or upgrading. This paper presents such a new methodology to construct mathematical programming models for energy and environment systems analysis. To acquire intended productivity and flexibility, the model is divided, at first, into three kinds of modules: a database system, a matrix generator and a report generator. Secondly, all the model elements are grouped into four categories and then the equations to express inter-element relationships are also categorized: "Flow" "Conversion Process", "Stock" and "Inter-regional Transportation". In order to verify the effectiveness of the new methodology, two kinds of sample models have been developed by the new methodology and their results are shown in this article. The new methodology helps attain high productivity in modeling systems of energy and environment and also of other areas. It also facilitates for the related people such as policy analysts or policy makers to interpret the model analysis results more accurately because the model structure is more clearly understood thanks to the categorization of model elements and equations.