A Method Combined Entropy with AHP Weighting Factors Affecting Safety on Construction Sites

摘要

The method of AHP often used in weight computation is a mathematical disposal of the expert’s subjective estimation. Weight is attained from judgment matrix which is evaluated by expert. So the factor weight is influenced by the subjectivity of expert. Appraisals from different expert are always different because of deficiency and discrepancy of each person’s experience and knowledge. Traditional method of processing those weights from different experts is weighting expert’s evaluation or geometric mean. They either have the character of subjectivity or accept the evaluations equally. If those evaluations are relatively concentrated, appraisals in concentrated area are more reliability, and if those evaluations are not relatively consistent, all of those evaluations should be considered. Here entropy, a measure of uncertainty in informatics, is introduced to measure the consistency of expert evaluations, combined with fuzzy set. In this way, subjective evaluation becomes more objective. In this paper, AHP and entropy are combined to evaluate hazard identification, and fuzzy set is also introduced. Through iterative algorithm of AHP analysis, the method considers fully sub-factors in construction safety hazard analysis, and it reduces relative subjectivity of expert, and distinguishes the important factors affecting safety on construction sites. it can give a more reliable factor weight as a reference for decision making.