Fuzzy Logic and Asset Allocation

Supervisors: Edy Portmann

Contact person: Edy Portmann

Student: Looking for student

Project status: Open

Year: 2025/2026

Thesis Scope

Asset allocation decision tools and procedures are predominantly in some form or shape versions of the well-known mean variance optimization. The mean-variance approach suffers from well documented deficiencies, however. These include high input parameter sensitivity or concentrated allocations. Practitioners typically don’t consider such outcomes “reasonable”. To overcome this hurdle various extensions and tweaks to the original framework have been proposed over time.

Instead of tweaking the original approach to cope with input uncertainty and the ambiguity of what reasonable means, one might consider an alternative approach all together. A technology that in principle by design should be well positioned to address these challenges is Fuzzy Logic.

Fuzzy logic technology has been around for a while. Only, how best to formulate the asset allocation problem so that it becomes suitable to be implement in a fuzzy system is not that obvious. The literature seems relatively sparse.

Thesis Goals

The goal this thesis is derive a suitable formulation of the asset allocation problem so that it can be processed by a fuzzy system. We broadly envision the project to follow three steps:

  1. Obtain a comprehensive understanding of the problems associated with the mean-variance asset allocation / portfolio construction approach.
  2. Obtain a comprehensive understanding of fuzzy logic and its applications. Specifically, rules-based inference and fuzzy optimization.
  3. Based on the above first two steps systematically evaluate the suitability of the different fuzzy methodologies to address the asset allocation problem.

Thesis Prerequisites

A successful candidate ideally has a background in asset allocation/portfolio construction as well as fuzzy logic. If not, the following knowhow should:

  • Be available be familiar with the terminology and concepts of financial markets such as asset classes, returns, correlation, volatility, portfolio theory.
  • Be familiar with optimization techniques such as quadratic programming, linear programming, convexity of optimization problems.

The project is a great opportunity for students working or intending to work in the asset management industry. It allows a deep insight into pros and cons of current methodologies and it allows to get exposure to a technology that gains renewed attention in the AI community.

Desired Skills (or willingness to learn): asset allocation, financial market theory, fuzzy design-science research.

Keywords: Asset Allocation; Fuzzy Sets; Fuzzy Logic; Linguistic Summarization.

Document: Not yet available