Optimal planning of energy management system under demand uncertainty

摘要

Energy management problem is a significant challenge as energy demand gets higher and securing resource is not always guaranteed. This study addresses the problem of modelling energy resource allocation and deriving on optimal policy for long-term investments. Model is constructed in Markov chain, which enables us to construct a probabilistic model that balances demand and supply using the data reflecting the current situation of South Korea. A large number of states with uncertainty make the resulting stochastic optimization problem near impossible to solve. We also propose an algorithmic strategy based on the framework of approximate dynamic programming and show this method is effective in solving the decision making problem.