Workshop Description
For production planners, operations research teams, and manufacturing technology leads. Covers quantum and quantum-inspired optimisation for job-shop scheduling, capacity planning, and multi-site production coordination. Includes QUBO formulations for manufacturing constraints, benchmark-specific performance comparisons against classical MILP solvers, and an independent vendor assessment.
Production scheduling is a combinatorial optimisation problem. Classical mixed-integer linear programming (MILP) solvers like CPLEX and Gurobi handle moderate instances well, but computational cost grows exponentially as you add machines, jobs, setup times, precedence constraints, and multi-site coupling. For a 200-job, 50-machine problem with realistic constraints, CPLEX can take hours to find near-optimal solutions. Quantum optimisation algorithms encode these problems as QUBO (Quadratic Unconstrained Binary Optimisation) and solve them using QAOA on gate-based hardware or quantum annealing on D-Wave systems. Published benchmarks from BMW, Airbus, and academic groups show that for certain scheduling structures (high constraint density, multiple objectives), quantum and quantum-inspired solvers find competitive solutions faster than classical approaches at problem sizes of 50 to 200 jobs. The question for manufacturing operations is whether this applies to their specific scheduling topology. This workshop maps that boundary through formulation exercises and performance comparisons relevant to delegates' own production environments.
What participants cover
- Classical scheduling limits: why MILP solvers hit computational walls on multi-machine, multi-objective production problems with real-world constraint density
- QAOA and quantum annealing for job-shop scheduling: how production constraints (machine availability, setup times, due dates, precedence) encode as QUBO
- Multi-site production coordination: coupling constraints between plants, shared resource pools, and inter-site transport as optimisation variables
- Benchmark evidence: published results from BMW, Airbus, and academic groups comparing quantum and quantum-inspired versus classical solvers at manufacturing-relevant problem sizes
- NISQ hardware limits: problem sizes tractable today (50-200 jobs on annealer, 20-50 on gate-based) and the fault-tolerant frontier for plant-scale scheduling
- Quantum-inspired alternatives available today: Fujitsu Digital Annealer, Toshiba SQBM+, and integration pathways with existing MES/ERP systems