Sectoral and spatial Benders methods speed multi-sector models

What the study found: The authors report that sectoral and spatial Benders decomposition methods, together with a budget-based formulation, improved the computational performance of multi-sector capacity expansion models. They say the algorithms produced runtime reductions of 15% to 70% compared with existing decomposition methods.

Why the authors say this matters: The study suggests these methods may help energy planning models remain computationally tractable while keeping high technological, spatial, and temporal resolution. The authors conclude that the approach can be applied to most existing energy planning models.

What the researchers tested: The researchers applied Benders decomposition, a way to split a large optimization problem into smaller linked parts, to multi-sector capacity expansion models. They developed sectoral and spatial decomposition algorithms and tested them on continental United States case studies with different spatial and temporal resolution settings.

What worked and what didn't: The reported algorithms outperformed existing decomposition algorithms in the tested cases, with runtime reductions between 15% and 70%. The abstract does not report any specific cases where the methods did not work.

What to keep in mind: The available summary does not describe detailed limitations beyond the tested case studies and configurations. The performance results are reported for continental United States case studies, so the abstract does not state how the methods perform in other settings.

Key points

• The study developed sectoral and spatial Benders decomposition algorithms for multi-sector capacity expansion models.
• A budget-based formulation was used to link the upper and sub-problems efficiently.
• The methods reduced runtime by 15% to 70% compared with existing decomposition methods.
• The test cases were based on the continental United States and varied spatial and temporal resolution.
• The abstract says the approach can be applied to most existing energy planning models.