“School of Cognitive Sciences”
Back to Papers HomeBack to Papers of School of Cognitive Sciences
| Paper IPM / Cognitive Sciences / 7535 |
|
||||||||||
| Abstract: | |||||||||||
|
Bounded rationality and satisficing models have shown good performance in reflection of uncertainties and complexities in real word problems of decision making and control. The emotional learning algorithm is such an appropriate method in multi objective applications. In this research the emotional learning algorithm is implemented on the base of the trainable Takaji Sugeno fuzzy inference system, which has shown great performance and flexibility among various learning methodologies. The two applications considered in this paper are selected among the well known real world problems: purposeful prediction of solar activity among the maximum regions, and multi objective portfolio selection. These multi criteria problems dealing with sampled data from unknown complex dynamical systems are hard to be handled with classical optimization techniques. In particular, the proposed emotional learning based fuzzy inference system (ELFIS) shows great performance in both cases: purposeful prediction and multi objective decision making. This method is also characterized by simple and fast computations which extends its use in large scale in real time systems. A comparison of the results with the adaptive network based fuzzy inference system (ANFIS) shows the efficiency of the proposed algorithm.
Download TeX format |
|||||||||||
| back to top | |||||||||||
