Complexity and Applicability

Complexity and Applicability

The complete machine-readable algorithm inventory is maintained in docs/algorithm-catalog.json and rendered automatically as Algorithm Catalog.

Reading the complexity table

• (T) is the number of planning periods.
• (p) is the number of worker threads/processors in the parallel complexity statement.
• "Worst case" means data-dependent pruning may be much faster on favorable instances.
• Memory figures refer to algorithmic working memory, excluding the immutable input and returned solution arrays.

Restricted methods

Several fast algorithms require one or more of:

• no speculative motive;
• stationary holding costs;
• constant unit production costs;
• strictly positive demand;
• nondecreasing setup costs.

The library does not silently coerce a general instance into a restricted model.

Numerical representation

Costs and quantities use double. Implementations check input validity and many accumulation paths explicitly guard against non-finite intermediate results.

For algorithm-by-algorithm details, open Algorithm Catalog and then the corresponding class or scientific page.