Unit commitment at industrial scale, benchmarked against the MILP you run
A unit-commitment problem of 25,755 variables and 48,939 constraints solved feasibly with a photonic quantum processor inside a hybrid MILP decomposition, on a national-laboratory testbed. The commercial MILP solver was better at short runtimes; the hybrid was competitive at longer ones. These findings do not establish a commercial quantum benefit.
Impact
The problem
Which generating units to commit, period by period, under ramping, reserve and network constraints. The problem is solved every day with a commercial MILP solver, and any new method has to be measured against it, not around it.
Our approach
The commitment sub-problem was carved out of a MILP decomposition and run on a photonic quantum processor; the rest stayed classical. Every quantum solution was checked for feasibility, and the runtime was compared with the commercial solver at matched quality.
The evidence
- 25,755 variables / 48,939 constraints; QPU solutions feasible; the commercial MILP solver better at short runtimes; the hybrid competitive at longer runtimes. Photonic processor plus classical decomposition.
- Pilot on a national-laboratory testbed. Next-generation figures in the source are projections and are not quoted here.
- Not claimed: a commercial quantum benefit. The source says so and this page carries it over.
Status: Platform network + Quantum hardware. Delivered through our platform network; photonic quantum processor plus classical MILP; pilot; the result does not establish a commercial quantum benefit.