report on ieee cis distinguished lecture program held …rcciit.org/docs/activity/a160218.pdf ·...
TRANSCRIPT
Report on IEEE CIS Distinguished Lecture Program held on 19th
November 2015 in the Department of Information Technology, RCC
Institute of Information Technology, Kolkata
The Department of Information Technology, RCC Institute of Information Technology,
Kolkata organized the IEEE CIS Distinguished Lecture Program on November 19th 2015 in
association with IEEE CIS Kolkata Chapter. Prof. Hisao Ishibuchi of Osaka Prefecture
University, Japan was the Distinguished Lecturer for the event.
The lecture started at Srinivas Ramanujam Auditorium of IT Department from 11:00 A.M.
Dr. Siddhartha Bhattacharyya, HoD (IT) and Dean (R & D), RCCIIT, Kolkata introduced to
the audience Prof. Hisao Ishibuchi. The audience comprised about 60 faculty members from
RCCIIT and other institutions. Prof. Arup K. Bhaumik, Principal, RCCIIT, felicitated Prof.
Ishibuchi with a bouquet of flowers.
The lecture was an enlightening one based on Evolutionary Many-Objective Optimization:
Search Behavior, Performance Indicators and Test Problems. It spanned across nearly 2
hours.
At the end of the lecture, Prof. Hisao Ishibuchi was handed over a token memento by Prof.
Arup K. Bhaumik. The lecture ended with a note of thanks to Prof. Hisao Ishibuchi by Dr.
Siddhartha Bhattacharyya.
Abstract of the lecture:
Evolutionary multi-objective optimization (EMO) has been an active research area in the
field of evolutionary computation for two decades. Recently various evolutionary many-
objective algorithms have been proposed in the literature. This is because many-objective
problems are difficult for frequently-used EMO algorithms such as NSGA-II and SPEA2.
This talk starts with a brief introduction to evolutionary computation. Next, some basic
concepts such as Pareto dominance, Pareto optimal solutions and Pareto front are explained
as an introduction to multi-objective optimization. After a basic framework of Pareto
dominance-based EMO algorithms is explained, co-evolutionary developments of test
problems and EMO algorithms are briefly explained. Then difficulties of many-objective
optimization are explained together with performance comparison results of some
representative EMO algorithms such as NSGA-II, MOEA/D, SMS-EMOA and HypE on
multi-objective and many-objective knapsack problems. It is also explained through
computational experimental results and simple numerical examples that performance
comparison results strongly depends on the specification of a reference point for the
hypervolume and a reference point set for the IGD indicator. Finally, solution set selection is
explained as a future research topics.