Parallelization of energy system optimization models using modified dual decomposition

Description

Linear programming (LP) is utilized by many energy system tools to optimize large scale models, but suffers from stagnating sequential computing power, further aggravated by continuously rising model size and complexity. Various approaches from optimization theory, like delayed column generation, exist to tackle the challenge of transforming given problems, in order to utilize parallel computing to solve models (faster). Of those, decomposition methods have been studied thoroughly for decades in operations research and energy related research, but are not commonly implemented in many energy system models.

Period	6 Sept 2022 → 9 Sept 2022
Event title	OR 2022
Event type	Other
Degree of Recognition	International