Parameter Tuning of Human Decision Making Model with Consideration of Human Factors through Metaheuristic Algorithms
Dr. Mitul Kumar Ahirwal
Maulana Azad National Institute of Technology, Bhopal (M.P.)
Assistant Professor
Dr. Mitul Kumar Ahirwal is currently working as an Assistant Professor in the Deptt. of Computer Science and Engineering, Maulana Azad National Institute of Technology, Bhopal. He has completed his Ph.D. from IIIT Jabalpur in 2014. He has more than 8 years of experience in teaching and research. His area of specialization includes Artificial Intelligence, Soft Computing, Evolutionary Intelligence, Biomedical Signal & Image Processing. He has published more than 50 research papers in SCOPUS/SCI indexed journals and International Conferences. He has published 03 books, Computational Intelligence and Biomedical Signal Processing (Springer), Artificial Intelligence Applications for Healthcare (Taylor & Francis) and Basics of UNIX Environment and System Calls (Edupublication). He is an active reviewer of many Elsevier, IEEE and Springer journals/conferences. He has completed 03 research projects, funded by Science and Engineering Research Board (SERB), Cognitive Science Research Initiative (CSRI), and SEED Grant by MANIT Bhopal. He has organized/conducted more than 10 STTPs/workshops. He is a senior member of IEEE, life member of International Association of Engineers (IAENG), Indian Academy of Neuroscience (IAN), and Soft computing Research society (SCRS) professional bodies. He has also been awarded Early Career Research Award /Grant (ECRA) by SERB in 2017.
Human decision-making (HDM) is a combination of various cognitive process, hence it is very difficult to model the process of HDM. Addition to this, different human factors affects decisions making (DM) process. Few of the well-known human factors are past experiences, emotions, times factors, and uncertainty. These factors are considered in the development computational model of HDM with the help reinforcement learning (RL). Iowa gambling Task (IGT) has been used for data collection. Final model consist of prospect utility, decay RL, and trial dependency choice rule (called as PU-DRI-TDC Model). The performance of developed model depends on several parameter like α, λ and C. For tuning of these parameters, metaheuristic optimization algorithm like Jaya and Real Coded Genetic Algorithm has been used. Mean Square Deviation (MSD) value is used to evaluate the model performance.