A Partial Order OWA Operator for Solving the OWA Weighing Dilemma

Prior weights are necessary for the application of ordered weighted averaging (OWA) operators, but obtaining them is expensive and contentious, which restricts the application of operators. To address the weighting issue, the weight space is used to “replace” the conventional weight vector, and the operator comparison is then extended to a partial order comparison on the weight space. The results show that the partial order OWA operator can be used as long as the weight order is clear, that is, there is no need to take accurate values. The evaluation result is represented by a Hasse diagram. The partial order OWA operator retains the properties of the conventional operator, and the running cost is low. It can be seen from the example that the partial order OWA operator solves the time weight problem. It can compare, sort, and optimize data using the Hasse graph, and it can also implement hierarchical clustering. The comparison results have strong robustness.