optiliste

Optiliste is a term used in optimization literature to denote a curated list of optimization methods organized by problem class and performance criteria. It serves as a decision-support resource to help practitioners select algorithms that are appropriate for a given mathematical model, data regime, and computational budget. The concept emphasizes comparability and transparency, presenting trade-offs between accuracy, speed, robustness, and resource use.

The word optiliste appears to be a compound of optimize and liste (list) found in several Germanic languages; it has surfaced in Danish, Norwegian, and some English-language writings to describe method catalogs. It is typically used in professional handbooks, course materials, and software documentation rather than formal theory papers.

Content and structure: An optiliste inventories problem classes such as linear programming, integer programming, convex optimization, nonconvex optimization, and combinatorial optimization, and maps them to algorithm families like exact solvers, decomposition methods, gradient-based methods, stochastic methods, and metaheuristics. For each pairing, it may provide guidance on expected performance, data requirements, and recommended parameter ranges, as well as representative benchmarks and typical failure modes.

Applications and use cases: Organizations use optilistes to select solvers in optimization pipelines, to benchmark new approaches, or to structure educational materials. They are also used in software libraries to document supported methods and their characteristics.

Limitations: The rapidly evolving nature of optimization means that a static optiliste can quickly become outdated; quality depends on the rigor of curation and the breadth of included problem classes. Users are advised to treat it as a starting point rather than a definitive ranking.

See also: Optimization, Benchmarking, Algorithm selection, Meta-learning.