A mathematical programming formulation for long-term infrastructure investment planning in Small Island Developing States

Abstract

Mixed-integer programming is a common method used in electricity generation and transmission optimization models. However, the size of the problem can result in extraordinarily long run times. Solve time also increases exponentially with the number of variables to optimize. There is therefore a constant trade-off between a realistic representation of the network and computational tractability. Additionally, actual data and publicly available, real-world application are scare. This is particularly true for Small Island Developing States. This paper bridges these gaps by describing a customized mathematical formulation for co-optimizing generation and transmission infrastructure investments. Data from the island of Jamaica and program scripts are available for reproduction. Key customizations to a mixed-integer programming model for long-term generation and transmission infrastructure investment planning include: • Hours are treated as representative hour categories and multiplied by the number of hour types within a given period. • Simulated construction is limited to every other year. • While fossil fuel plants are treated as discrete variables, renewable energy plants are treated as continuous variables.

Keywords

• Small island developing states
• Energy modeling
• Generation capacity
• Transmission capacity
• Co-optimization