Exact Algorithms

Exact Algorithms

An exact strategy returns an optimal ULS solution when its documented applicability conditions are satisfied.

The direct exact methods are grouped by mechanism rather than by release:

Dynamic programming and geometric acceleration

Wagner–Whitin, Evans, Wagelmans, Federgruen–Tzur, Aggarwal–Park and related linear or O(T log T) variants.

Planning-horizon methods

Bahl–Taj, Heady–Zhu and Sadjadi–Aryanezhad–Sadeghi use data-dependent pruning or horizon reduction while preserving exactness under their stated assumptions.

Parallel methods

LyuLeeParallelSolver exposes the parallel literature line as a separate public implementation.

Network and combinatorial methods

ZangwillNetworkSolver uses a shortest-path interpretation. JacobsKhumawalaBranchAndBoundSolver represents the branch-and-bound literature line.

Solver-backed exact strategies

Mathematical formulations and (l,S) cutting-plane solvers are exact too, but they are shown separately in the user documentation because they depend on an optimization engine.

For a method-by-method comparison, open Algorithm Catalog or use the card-based documentation portal.