Novel competitive-cooperative learning models (cclms) based on higher order information sets
- PDF / 1,492,050 Bytes
- 18 Pages / 595.224 x 790.955 pts Page_size
- 102 Downloads / 152 Views
Novel competitive-cooperative learning models (cclms) based on higher order information sets Jyotsana Grover1 · Madasu Hanmandlu2
© Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract This paper presents two novel competitive-cooperative learning models (CCLM) for achieving goals by human contenders. These models have two phases, viz., Competition phase and Cooperation phase. CCLM based on Hanman Transform (HT) is called HT-CCLM and that using a new concept termed Pervasive Information set is called PIS-CCLM. In the competition phase of HT-CCLM, each contender emulates the effort of best achiever by taking the difference of Hanman Transform values associated with the efforts of an individual and the best achiever whereas in the cooperation phase the differential of HT values of efforts of two random contenders is considered. In PIS-CCLM pervasive information value obtained from hesitancy values are used in the competition phase only. We have also carried out Wilcoxon test to establish the superiority of the proposed HT-CCLM and PIS-CCLM. Keywords Hanman transform · Hesitancy · CCLM · HEFAG · Entropification · Competition and cooperation phases · Entropy function and hesitancy degree · PIS
1 Introduction In the recent times meta-heuristic learning techniques have overtaken the classical optimization methods that require derivatives of an objective function and suffer from both computation and convergence problems if the number of parameters is very large. The meta-heuristic learning methods being population based are not crucially dependent on the initial parameters values unlike the classical methods each of which yields one solution whereas the
Jyotsana Grover is in Department of Computer Science and Information Systems WILP Delhi Center, BITS Pilani, India, email:[email protected] Jyotsana Grover
[email protected] Madasu Hanmandlu hanmandlu [email protected] 1
Department of Computer Science and Information Systems WILP Delhi Center, BITS Pilani, Pilani, India
2
CSE Department, MVSR Engg. College, Nadergul, Hyderabad, 501510, India
meta-heuristic techniques explore the solution space to look for the optimal solution through the agents that adopt their respective strategies for achieving their respective goals. We present here a novel meta-heuristic method named as Competitive Cooperation Learning Model (CCLM) for achieving a goal with the help of effort in two phases: Competition and Cooperation. Our previous work, Human Effort For Achieving Goal (HEFAG) [1] also aims at achieving a goal in two phases: Emulation and Boosting. But the present work differs from the previous work in several ways: i) HEFAG employs the basic information values whereas CCLM employs higher level information values, ii) The emulation phase involves three membership functions while the competition phase two, iii)Incorporating reinforced learning in competition phase is very easy through Hanman Transform and hesitancy but it is not so in emulation phase, and iv
Data Loading...
