An algorithm for automatic assignment of reviewers to papers

The assignment of reviewers to papers is one of the most important and challenging tasks in organizing scientific conferences and a peer review process in general. It is a typical example of an optimization task where limited resources (reviewers) should be assigned to a number of consumers (papers), so that every paper should be evaluated by highly competent, in its subject domain, reviewers while maintaining a workload balancing of the reviewers. This article suggests a heuristic algorithm for automatic assignment of reviewers to papers that achieves accuracy of about 98–99% in comparison to the maximum-weighted matching (the most accurate) algorithms, but has better time complexity of Θ(n2). The algorithm provides an uniform distribution of papers to reviewers (i.e. all reviewers evaluate roughly the same number of papers); guarantees that if there is at least one reviewer competent to evaluate a paper, then the paper will have a reviewer assigned to it; and allows iterative and interactive execution that could further increase accuracy and enables subsequent reassignments. Both accuracy and time complexity are experimentally confirmed by performing a large number of experiments and proper statistical analyses. Although it is initially designed to assign reviewers to papers, the algorithm is universal and could be successfully implemented in other subject domains, where assignment or matching is necessary. For example: assigning resources to consumers, tasks to persons, matching men and women on dating web sites, grouping documents in digital libraries and others.