Quantum lab
The pricing decision is written as a QUBO — one binary variable per product-and-price pair — and solved with quantum-inspired annealing. Every run is benchmarked against an exact classical solve so you can trust the answer.
Problem formulation
Binary price-selection variables with hard and soft constraints.
Variables. x[p,k] = 1 when product p is sold at candidate price k (five candidates per product, 0.90× to 1.10× of today's price).
One price per product. Enforced as a one-hot penalty A·(Σ x[p,k] − 1)² and preserved structurally by the annealer's move set.
Minimum margin. Candidates below the margin floor are removed before the QUBO is built, so infeasible prices cannot be selected.
Price stability. A quadratic penalty pulls the average price index toward 1.00 and a band penalty keeps it inside 0.98–1.02.
Objective. Maximise weighted profit, inventory health and customer value; subtract waste and price movement.
Annealing energy per restart
Lower energy is a better solution.
Benchmark across runs
Gap in percent against exact solve time.
Bubble size is annealing time. A gap of 0% means the quantum-inspired solver reproduced the exact optimum.