Solutions
From TSP and job-shop scheduling to portfolio construction and drug discovery — GPU-accelerated QUBO solvers for any combinatorial structure.
Problem types
Exact and heuristic solvers for symmetric and asymmetric TSP. GPU annealing returns < 0.5% gap on TSPLIB-1002 in under 4 seconds.
CVRP, VRPTW, multi-depot, and pickup-delivery variants. Constraints encoded as quadratic penalties.
Mean-variance QUBO formulation with cardinality constraints, sector bounds, and index-tracking objectives.
Classic and flexible job-shop, open-shop, and staff scheduling. Tabu Search exploits operation-swap neighbourhood.
Direct QUBO mapping — no penalty overhead. QAOA achieves ≥ 0.6924 approximation ratio at p = 1 on 3-regular graphs.
1D and 2D bin packing, knapsack, and cutting stock. Compact penalty-based QUBO encoding.
Submit any Q matrix directly — dense numpy, scipy COO/CSR, or JSON. Full solver access, result includes gap_to_best.
By industry
Last-mile routing, fleet scheduling, warehouse slotting, and cross-dock optimisation at scale.
Portfolio construction, risk decomposition, derivatives hedging, and index replication under cardinality constraints.
Flexible job-shop scheduling, production sequencing, bin-packing for packaging, and defect clustering.
Facility location, community detection, graph partitioning, and minimum-cost coverage problems.
Supplier selection, inventory allocation, order batching, and demand-constrained routing.
Molecular docking poses, protein contact-map folding, and binding-site MaxCut formulations.
GPU-native quantum-inspired optimization for combinatorial problems at any scale.