Evolutionary multi-objective optimization and its application in games and robotics

Kalyanmoy Deb
Deva Raj Chair Professor
Dept. of Mechanical Engineering
Indian Institute of Technology Kanpur, 
208016, India

Date: 15 September 2010
Time: 9 - 10 am

Abstract: Multi-objective optimization deals with multiple conflicting objectives which must be considered simultaneously. In principle, such problems give rise to a set of trade-off Pareto-optimal solutions. The advent of evolutionary multi-objective optimization (EMO) for the past two decades is motivated mainly due to the population approach associated with evolutionary algorithms. In this talk, we shall discuss the principles of EMO and present an efficient algorithm for solving two and three-objective problems. One of the popular extensions is that the principle of EMO is used to solve other problems which may not be multi-objective in the first place. By considering additional objectives into the optimization process, more information about the problem and better solutions are possible to be found. As an illustration, we shall discuss two problems -- iterated prisoner's dilemma problem and a robot path navigation problem. In both problems, we shall demonstrate how an EMO can be used to solve such problems better than usual single-objective methods.


Dr Kalyanmoy Deb (Phd in Engineering Mechanics from the University of Alabama (Tuscaloosa)) is a Professor of Mechanical Engineering (Indian Institute Of Technology Kanpur) and the Director of the Kanpur Genetic Algorithms Laboratory (KanGAL) which he established in 1997. 

Prof. Deb received his Bachelor's degree from 
IIT Kharagpur (Mechanical Engg 1985). Before joining Alabama, Prof. Deb served with Engineers India Limited (New Delhi) between 1985 and 1987. He was also a Visiting Research Assistant Professor in the Department of General Engineering at the University of Illinois, Urbana Champaign between 1991 and 1992 and worked at Illinois Genetic Algorithms Laboratory (IlliGAL)

Recipient of Outstanding Graduate Research Assistant award for 2 consecutive years at Alabama, Prof. Deb has numerous awards and honours in his name, including the prestigeous Shanti Swarup Bhatnagar Prize in Engineering Sciences in 2005, `Thomson Citation Laureate Award', an award given to an Indian Researcher for making most highly cited research contribution during 1996-2005 in a particular discipline according to ISI Web of Science.,Friedrich Wilhelm Bessel Research Award and Humboldt Fellowship from Alexander von Humboldt Foundation, Germany. He is a fellow of 
Indian National Academy of Engineering (INAE)Indian Academy of Sciences (IASc), andInternational Society of Genetic and Evolutionary Computation (ISGEC). At present Prof. Deb is an Associate Editor of the following international journals (i) IEEE Transactions on Evolutionary Computation Journal and (ii) Evolutionary Computation Journal from MIT Press , an Editorial Board Member of (i) Engineering Optimization Journal and (ii)Genetic Programming and Evolvable Machines . He has also the rare honour of being the only Asian Executive Council Member of the International Society for Genetic and Evolutionary Computation (ISGEC) 

Author of more than 240 research papers and two textbooks, his latest book on 
Evolutionary Multiobjective Optimization Algorithms is the first ever compilation of multiobjective optimization algorithms. Because of his pioneering research in the field of evolutionary multi-objective optimization, he has been invited to present 26 Keynote lectures and more than 100 invited lectures and tutorials on the topic. Professor Deb has organized several conferences and founder-chaired the First Conference on Evolutionary Multicriterion Optimization (EMO 2001) held at Zurich. 

His research has a practical bend, because of which many researchers and applicationists refer to his research. His NSGA-II paper from IEEE Trans. on Evolutionary Computation (2000) is judged as the 
Fast-Breaking Paper in Engineeringby ESI Web of Science recently. Immensely popular among his students and colleagues, Prof. Deb happens to be the most cited Indian researcher in the Computer Science area according to NEC CiteSeer Research Index. Citation Count as per Web of Science : over 9000 ISI citations since 1993 (counted in December 2008).

Evolutionary Multi-objective Footstep Planning for Humanoid Robots

Jong-Hwan Kim
Professor, IEEE Fellow, FIRA President
Dept. of EE, Korea Advanced Institute of Technology (KAIST)

Date: 16 September 2010
Time: 9 - 10 am

Abstract: This talk presents a novel evolutionary multi-objective footstep planner for a humanoid robot. Firstly, a footstep planner using univector field navigation method is introduced to provide a command state (CS) which is to be an input of modifiable walking pattern generator (MWPG) at each footstep. Then MWPG generates corresponding trajectories of every leg joint of the humanoid robot at each footstep to follow the CS. Secondly, multi-objective evolutionary algorithm (MOEA) is employed to optimize univector fields satisfying multiple objectives in navigation. Lastly, preference-based selection algorithm based on fuzzy measure and fuzzy integral is presented to select a preferred one out of various nondominated solutions obtained by MOEA. The effectiveness of the proposed evolutionary multi-objective footstep planner is demonstrated through computer simulations for a simulation model of the small-sized humanoid robot, HanSaRam-VIII (HSR-VIII).

Biography: Jong-Hwan Kim joined the Department of Electrical Engineering and Computer Science, KAIST, Korea in February 1988 and is currently a full professor. Dr. Kim established two national robotics research centers in 2000 and 2003, respectively, as Director and has been serving as Director for KT RoboLab@KAIST since 2007. His research interests include evolutionary algorithm, computational intelligence, evolvable artificial creature, soccer robotics, ubiquitous and genetic robotics and cyber-physical robotic system. Dr. Kim has authored 5 books and 2 edited books, 2 journal special issues and around 300 book chapters and refereed papers in technical journals and conference proceedings and his current h-index is 25 along with over 2,000 number of citations. He has delivered over 130 invited talks including 32 keynote speeches at the international conferences in 21 countries. Dr. Kim was an advisory professor for the SBS television drama, ‘KAIST’, which was aired over one year starting in Jan. 1999. His activities have been introduced as news items and articles in world media such as CNN, BBC, TF1, The Times, etc. and as documentaries four times in Korea broadcasting companies including KBS, MBC, etc. ( Dr. Kim currently serves as an Associate Editor of the IEEE Transactions on Evolutionary Computation, the IEEE Computational Intelligence Magazine and the International Journal of Social Robotics. He was one of the co-founders of the International Conference on Simulated Evolution and Learning (SEAL) in 1996. He was General Chair for the IEEE Congress on Evolutionary Computation, Seoul, Korea, 2001 and is General Chair for the IEEE International Symposium on Computational Intelligence in Robotics and Automation, Jeju Island, Korea, 2009 and Honorary Chair for FIRA RoboWorld Congress, Incheon, Korea, 2009. He has been on the program committees and advisory boards of more than 100 international conferences and co-chaired several international conferences. His name was included in the Barons 500 Leaders for the New Century in 2000 as the Father of Robot Football. Dr. Kim is the Founder of FIRA (The Federation of International Robosoccer Association) and IROC (The International Robot Olympiad Committee). He is currently serving FIRA and IROC as President. Dr. Kim was the recipient of the science and technology award from the President of Republic of Korea in 1997 and has been elevated to 2009 IEEE Fellow.