A COMPREHENSIVE SURVEY OF THE REVIEWER ASSIGNMENT PROBLEM

Reviewer Assignment Problem (RAP) is an important issue in peer-review of academic writing. This issue directly influences the quality of the publication and as such is the brickwork of scientific authentication. Due to the obvious limitations of manual assignment, automatic approaches for RAP is in demand. In this paper, we conduct a survey on those automatic approaches appeared in academic literatures. In this paper, regardless of the way reviewer assignment is structured, we formally divide the RAP into three phases: reviewer candidate search, matching degree computation, and assignment optimization. We find that current research mainly focus on one or two phases, but obviously, these three phases are correlative. For each phase, we describe and classify the main issues and methods for addressing them. Methodologies in these three phases have been developed in a variety of research disciplines, including information retrieval, artificial intelligence, operations research, etc. Naturally, we categorize different approaches by these disciplines and provide comments on their advantages and limitations. With an emphasis on identifying the gaps between current approaches and the practical needs, we point out the potential future research opportunities, including integrated optimization, online optimization, etc.