Finite Memory and Dynamic Decision Problems

2026-07-22

Memory limitations lie at the core of how agents store past information and make decisions. For example, consider a buyer choosing between two alternative brands of a good. In thinking about a brand, she may recall her past experiences, product reviews, or expert’s ratings, which induce a belief about the brand’s quality. In this process, the buyer may divide the quality scale into finitely many categories such as high, medium, and low. Then, she may switch from one brand to another only if her belief about a brand’s quality moves from one category to another. Although memory limitations in the form of finite memory naturally arise in many facets of economic life, formal economic models that incorporate the finite memory restriction are rather limited. The aim of this project is to understand the implications of finite memory for dynamic decision problems in both choice theoretical and strategic economic contexts. These implications are significant to reach out a wide range of objectives such as obtaining better predictions of consumer’s choices, unravelling the observed equilibrium behaviour, and regulating markets as to maximize social welfare. In addition, the results to be obtained can provide new explanations for observed economic phenomena such as limited attention and price stickiness. The outgoing phase of this GF project is divided into two parts. This first part of the project aims at addressing a fundamental problem to expedite the use of finite memory model in economic applications. In the second part, I aim to examine how strategic pricing aspects of a dynamic duopoly (oligopoly) model. The goal in here is to incorporate memory limitations in the form of finite memory into a fundamental economic problem, and identify its effects.This GF project will be carried out in Princeton University under the supervision of Prof. Faruk Gul. I will then return to Bilkent University to improve my results for 12 months under the supervision of Prof. Semih Koray.

Turkey

2026-02-01