A Methodology for Aggregation of Multiple Criteria Rank-Ordered Priorities

摘要

This report is concerned with the aggregation of multiple rank-ordered research and development projects or product requirements. The resultant prioritized list serves as a resource allocation basis for the R&D projects. The rank orders are ordinal and without feedback or strategy. The extensive literature of the social choice field is structured and analyzed for this research. Building upon that foundation, this research classified the known majority-rule aggregation methods, evaluated them, and developed a model and computer code using the better methods to compute the aggregation of up to 100 rank-ordered lists of up to 100 alternatives. Decision-maker and alternative weighting may be applied by any of eight methods. Judge self-evaluation weighting is also available. Final rank-ordered priority lists of R&D projects are computed by the Borda, Adjusted Borda, and Shannon Majority-Rule methods. The preferred Shannon method handles complete, partial, and weighted rank orders without transitivity. Fuzzy set rank orders are compared. Kendall's coefficients of consistency and concordance evaluate the rank orders. The model is validated by comparative computation of many rank-ordered aggregation examples from the literature. (Author)