Full reinforcement operators in aggregation techniques

摘要

We introduce the concept of upward reinforcement in aggregation asnone in which a collection of high scores can reinforce or corroborateneach other to give an even higher score than any of the individualnarguments. The concept of downward reinforcement is also introduced asnone in which low scores reinforce each other. Our concern is with fullnreinforcement aggregation operators, those exhibiting both upward andndownward reinforcement. It is shown that the t-norm and t-conormnoperators are not full reinforcement operators. A class of operatorsncalled fixed identity MICA operators are shown to exhibit the propertynof full reinforcement. We present some families of these operators. Wenuse the fuzzy system modeling technique to provide further examples ofnthese operators