This work concerns resource allocation in a power grid. Scheduling of generation by a central authority takes place in two stages:
- Unit commitment: Which generators are running and “plugged in”
- Level of generation for each generator.
The first decision is cast as a combinatorial optimization problem, which is generally considered difficult by electrical engineer/computer science/computational academics
The second step can be cast as a convex optimization problem, and most in these academic communities would declare that this problem is easy. Whey then do the ISO/RTOs in the U.S. find step 2 so difficult?
One reason is that the objective function is not very sensitive to the precise allocation, but the market demands precise solutions to the optimization problem. Slight errors in computation can mean great rewards to some participants in the energy market. Consider the 2013 finding regarding Deutsche Bank’s alleged manipulation of energy markets, (Reuters) – German bank Deutsche Bank AG will pay nearly $1.7 million to settle allegations it manipulated electricity markets in California in 2010
These papers introduce more efficient and reliable computation tools for computation of allocations in energy markets to address these points.
References
@article{wanshazhelitmey13a,
Title = {An Extreme-Point Subdifferential Method for Convex Hull Pricing in Energy and Reserve Markets. {Part I:} Algorithm Structure},
Author = {Wang, Gui and Shanbhag, U.V. and Zheng, Tongxin and Litvinov, E. and Meyn, S.},
Journal = {{IEEE Transactions on Power Systems}},
Number = {3},
Pages = {2111-2120},
Volume = {28},
Year = {2013}}
@article{wanshazhelitmey13b,
Title = {An Extreme-Point Subdifferential Method for Convex Hull Pricing in Energy and Reserve Markets. {Part II:} Convergence Analysis and Numerical Performance},
Author = {Wang, Gui and Shanbhag, U.V. and Zheng, Tongxin and Litvinov, E. and Meyn, S.},
Journal = {{IEEE Transactions on Power Systems}},
Number = {3},
Pages = {2121-2127},
Volume = {28},
Year = {2013}}