Four new ordered weighted averaging weights generators for regular increasingly monotonic functions

Abstract

Diverse normalized weight vectors for OWA aggregation can be generated using regular increasing monotonic functions, embodying bipolar optimism–pessimism preferences. Yager’s original approach has been utilized for over three decades. This work, from different perspectives, proposes and analyzes four approaches to generate weight vectors with regularly increasing monotonic functions. We systematically formulate and analyze Yager’s original method and formally define it as a generator. Furthermore, we propose and analyze four distinct generators with different features and characteristics. The first two possess attenuation properties compared to Yager’s generator. The third one offers a significant advantage of full consistency in orness, and the fourth one provides a consistent cognitive mode for the regularly increasing monotonic functions that coincide almost everywhere.