Current events

Algorithms that decide on investments

  Foto: Unsplash/Austin Distel

Foto: Unsplash/Austin Distel

UOC researchers promote artificial intelligence to improve economic decision making of companies and governments

Machines have been performing tasks and solving problems for many years; however, there are decisions that entail risks that still depend to a large degree on humans. In the case of fields such as finance, the path chosen may make the difference between success and failure. To help in this kind of decision, the Universitat Oberta de Catalunya (UOC) has entered into an international collaboration to develop algorithms that will automate risk assessment and decision-making for investments.

In the context of what is known as computational finance, UOC researchers Jana Doering and ngel Juan, from the Internet Interdisciplinary Institute's (IN3) ICSO research group, together with researcher ngels Fit, from Management & eLearning (MeL) research group of the Faculty of Economics and Business, have published a paper, in the journal Operations Research Perspectives, on the implementation of metaheuristic algorithms in finance and how they can be used to solve highly complex problems in the field of computing and combinatorial optimization. Specifically, the experts address the optimization of investment portfolios and management of the associated risks. The research team has analysed the relationship between these two areas and proposes new hybrid algorithms, such as the so-called simheuristic (a combination of metaheuristics and simulation) and learnheuristic (a combination of metaheuristics and machine learning) algorithms to generate solutions that significantly improve decision-making in financial environments.

 

Automated finance to avoid human errors

"Research in this field has grown significantly in recent years and, although some metaheuristics have received more attention than others, different implementations have found solutions to priority problems for financial organizations," said Doering. "In the financial world, we find that many complexities of real life are not factored in. So, it may be useful to combine metaheuristics with machine learning to model behavioural dynamics and with simulation to define uncertain parameters."

As decisions are made applying smarter criteria based on the use of data and latest-generation algorithms, the researchers believe that their efficiency will increase considerably. "The idea is that small and large investors alike could benefit from an optimization algorithm that helps them make successful decisions in an asset portfolio, in line with their yield preferences," said ngels Fit, who also is UOC Vice President for Competitiveness and Employability. The UOC experts added that it would be good to have an independent platform for deciding on optimal investments with a limited risk. This type of risk management analysis could also be useful for any institution or government when it has to make decisions, particularly those that involve a budget investment.

 

Multiple uses

The ICSO team are experts in data analytics and the development of metaheuristics, that is, smart algorithms that help optimize decision-making in many other spheres beyond finance. "Using algorithms, we can design better transport systems and more efficient telecommunication networks, improve energy-saving processes, and define more sustainable mobility policies in smart cities, insurance risks or, precisely, high-yield financial strategies," said ngel A. Juan, UOC professor, co-author of the study and the group's lead researcher.

This research project has received financial support from the Divina Pastora Seguros insurance company, and input from researchers Renatas Kizys, from the University of Southampton, United Kingdom, and Onur Polat, from Bilecik Şeyh Edebali University, Turkey.

 

Reference article

Doering, J.; Kizys, R.; Juan, A.; Fit, A. and Polat, O. (2019). "Metaheuristics for rich portfolio optimisation and risk management: Current state and future trends". Operations Research Perspectives. doi: <https://doi.org/10.1016/j.orp.2019.100121>

 

UOC R&I

The UOC's research and innovation contributes to overcoming the challenges faced by the global societies of the 21st century. It focuses on studying the interaction between human activity and information and communication technologies, paying particular attention to e-learning and digital health. The UOC's more than 400 researchers and 46 research groups are linked to its seven faculties and three research centres: IN3, eLearn Center and eHealth Center.

The UN's 2030 Agenda Sustainable Development Goals for a fairer, more equitable world and open knowledge mark the strategic bases for the University's teaching, research and innovation. More information: research.uoc.edu.

#UOCexperts

ngel Juan

Full professor and ICSO research group leader, IN3

Expert in: Optimization and simulation algorithms applied to logistics, transport and manufacturing.

Knowledge area:

Quantitative and computational methods applied to business.

Jana Doering

Researcher of the ICSO, IN3 group

Expert in: Optimization and simulation algorithms applied to logistics, transport and manufacturing.

Knowledge area:

Quantitative and computational methods applied to business.

Photograph of ngels Fit Bertran

ngels Fit Bertran

Expert in: Business management systems, financial analysis, business ethics, and management and e-learning.

Knowledge area: Accounting, management accounting and e-learning.

View file