A stochastic programming model for an energy planning problem: formulation, solution method and application

• Published in: Annals of operations research Vol. 311; no. 2; pp. 695 - 730
• Main Authors: Irawan, Chandra Ade, Hofman, Peter S., Chan, Hing Kai, Paulraj, Antony
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.04.2022; Springer; Springer Nature B.V
• Subjects: Business and Management; Carbon dioxide; Combinatorics; Computing time; Electric power demand; Energy management systems; Industrial plant emissions; Methods; Operations research; Operations Research/Decision Theory; Power plants; Production methods; Regional planning; S.i.: Mopgp19; Stochastic models; Stochastic programming; Theory of Computation; China; Multi-objective optimization; Energy planning; Stochastic programming
• ISSN: 0254-5330; 1572-9338
• DOI: 10.1007/s10479-020-03904-1

The paper investigates national/regional power generation expansion planning for medium/long-term analysis in the presence of electricity demand uncertainty. A two-stage stochastic programming is designed to determine the optimal mix of energy supply sources with the aim to minimise the expected total cost of electricity generation considering the total carbon dioxide emissions produced by the power plants. Compared to models available in the extant literature, the proposed stochastic generation expansion model is constructed based on sets of feasible slots (schedules) of existing and potential power plants. To reduce the total emissions produced, two approaches are applied where the first one is performed by introducing emission costs to penalise the total emissions produced. The second approach transforms the stochastic model into a multi-objective problem using the ϵ -constraint method for producing the Pareto optimal solutions. As the proposed stochastic energy problem is challenging to solve, a technique that decomposes the problem into a set of smaller problems is designed to obtain good solutions within an acceptable computational time. The practical use of the proposed model has been assessed through application to the regional power system in Indonesia. The computational experiments show that the proposed methodology runs well and the results of the model may also be used to provide directions/guidance for Indonesian government on which power plants/technologies are most feasible to be built in the future.