4 Objective Problem Test Cases for Knapsack, Assignment, and TSP Algorithms

A set of randomly generated test instances for three classic multi-objective optimization problems: knapsack, assignment, and traveling salesman. The dataset was created by William Pettersson using generation techniques from established academic papers. The knapsack instances have item weights and objective coefficients drawn uniformly from ranges like {60,...,100}, with a capacity bound set to half the total item weight.

Use Cases

• Benchmarking multi-objective knapsack solvers based on randomly generated integer weights and objective coefficients.
• Testing assignment problem algorithms using instances with objective coefficients drawn from {0,...,20}.
• Evaluating traveling salesman problem solvers on both Euclidean instances with integer coordinates and non-Euclidean instances with random distances.
• Comparing algorithm performance on problems with multiple objectives defined in an extended LP file format.

Strengths

• Test cases are generated using methods from three peer-reviewed publications (Laumanns et al. 2006, Przybylski et al. 2010, Özpeynirci and Köksalan 2010).
• Includes both Euclidean and non-Euclidean traveling salesman problem variants, with distances drawn from {20,...,180}.
• Knapsack problem parameters are clearly defined, with item weights and objective coefficients drawn from specific integer ranges.

Limitations

• The exact number of instances, rows, and file size is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred from the description or after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

William Pettersson via paperswithcode.

Collection Method

Random generation following techniques from specific academic papers for each problem type.