Studies in Classical Behavioural Economics: A Thesis Proposal

I propose to investigate the computable foundations of decision making along the lines of Classical Behavioural Economics. First, decision makers will be represented as Information Processing Systems underpinned by bounded rationality and satisficing as in the theory of Human Problem Solving proposed by Herbert Simon. The game of Go, which has a finite and large search space, provides a rich paradigm for analyzing how boundedly rational individuals use “heuristics” (algorithm) to solve highly complex problems. Hence, I will use Go to study different aspects of human problem solving, via simulations. Second, I will make a case for posing human decision problems as Diophantine decision problems, as opposed to optimization problems (as in Modern Behavioral Economics) and viewing full rationality as a special case of bounded rationality. Finally, the role of Near Decomposability, a recurring theme in Simon’s works, and its relevance in modeling human behavior and learning will be explored.