Problem Classes
The Intractable Decathlon — ten challenging optimization problem classes. Jump straight to any class below, or browse the cards for a summary of how far each instance family has been solved.
These instances stress multi-constraint subset-sum structure, where feasibility is easy to state but difficult to certify at useful sizes.
LABS is a canonical spin benchmark with direct links to communications, radar, and cryptography, and it becomes harder as sequence length grows.
Minimum Birkhoff decomposition links assignment structure, sparse representation, and quantum physics applications through a hard cardinality objective.
Steiner tree packing models wire-routing pressure in VLSI-style grids, where many connection demands must coexist without conflicts.
Sports timetabling captures realistic constraint interactions from round-robin tournaments, with instances selected for diversity and difficulty.
Portfolio instances add transaction costs, short selling, borrowing costs, and time coupling to a familiar financial optimization model.
Maximum independent set is a fundamental graph problem with compact QUBO structure and hard instances from social, biological, and benchmark graphs.
Network design represents traffic-routing and degree-constrained infrastructure planning, with objective values tied to congestion.
Vehicle routing combines route selection, capacity, and time-window pressure, reflecting core logistics and mobility applications.
Topology design asks for low-diameter graphs under degree limits, a concise model for communication latency and network architecture.