RandomCsp is a set of C++ classes and a set of programs for generating test sets for binary constraint satisfaction problems. Several models are currently available. The classes are useful for implementing algorithms that solve these problems.
|Tags||Scientific/Engineering Artificial Intelligence Software Development Testing|
Release Notes: This release adds the model RB by Xu and Li. It works with gcc 3.3.2. The default output has been changed from list to matrix.
Release Notes: This release adds Model F which is an improvement over Model E when concerned with the density property in constraint satisfaction problems.
Release Notes: This release fixes a bug in RemoveConflicts that caused problems with the counting of the number of conflicts.
Release Notes: A new file format that potentially saves disk space was added. A new application to analyze CSP instances was added. A new generation model, called Model G, which generates instances that prove difficult to solve for evolutionary algorithms was added.
Release Notes: A new random model (model G) was added that creates solvable instances that are deceptive for evolutionary algorithms. A bug in model E was fixed, creating instances that were too easy. Older models were renamed to appropriate names (model A and model B).