testknap

Testknap is a software toolkit designed to support research and education in the knapsack problem by providing standardized benchmarks and evaluation tools. It aims to enable fair and reproducible comparisons of knapsack solvers across variants, instance sets, and problem scales.

The toolkit comprises four main components: an instance generator capable of producing a wide range of knapsack instances (0/1 knapsack, bounded, unbounded, and fractional variants), a solver interface that lets users integrate exact methods and heuristics, a benchmarking engine that runs solvers under controlled conditions, and an evaluation/reporting module that computes metrics such as best objective value, optimality gap, runtime, and memory consumption.

Instances and results are stored in standardized formats to support reproducibility and interoperability, with support for CSV, JSON, and common knapsack instance formats. The project emphasizes reproducibility by capturing solver configurations and environment details alongside results.

Applications include benchmarking new algorithms, validating theoretical analyses, and supporting classroom demonstrations. It is used by researchers, educators, and software developers working on combinatorial optimization and related areas.

Testknap is an open-source project with an active community of contributors. It provides documentation, example datasets, and tutorials to help new users adopt the toolkit for research or teaching.