DESDEO: The Open-Source Software Framework for Interactive Multiobjective Optimization

Many real-world decisions—especially in resource-constrained settings such as developing countries—require balancing several conflicting objectives at once, whether trading off cost against sustainability or growth against equity. Interactive multiobjective optimization methods address this by engaging the decision maker throughout the solution process, progressively narrowing in on a satisfactory compromise rather than requiring the costly generation and interpretation of an entire Pareto frontier. This makes such methods particularly practical where computational resources and time for exhaustive analysis are limited, yet informed trade-off decisions are still essential.

Developing countries, like other countries, face challenges in

Land and natural resources usage—balancing the needs of economic activity with the needs of the local population, and environmental protection Energy system planning—balancing investment costs, reliability, price of electricity, and environmental impact Emergency preparedness—balancing up-front costs and the ability to get the aid to the people impacted by, for example, a storm in a timely manner.

By treating these challenges as multiobjective problems they are, the responsibility of balancing the conflicting objectives is no longer in the hands of the OR practitioner. Interactive multiobjective optimization methods are designed to let the decision maker explore the different possible solutions and then make an informed decision that balances the conflicting goals.

DESDEO (found at https://github.com/industrial-optimization-group/DESDEO) is an open-source framework designed to put these interactive methods into practice. It supports modelling and solving multiobjective optimization problems using multiple different methods that can be powered by several different free or commercial optimization solvers or evolutionary optimization methods. DESDEO is composed of three major components:

The core-logic, which contains the algorithmic implementation of interactive methods, various tools related to multiobjective optimization and means to model a variety of multiobjective optimization problems. The core-logic can be considered stable enough for use in research and applications. The web-API (WIP), which implements a web-based application programming interface (API) to allow the use of the various functionalities found in DESDEO's core-logic through a web connection. The web-API also implements a database, which is a vital component for managing and enabling decision-support using the framework.

The web-GUI (WIP), which implements a web-based interface for utilizing the interactive methods and tools for modeling and solving multiobjective optimization problems. There is documentation (https://desdeo.readthedocs.io/en/latest/) that is also being continuously developed as DESDEO evolves and it contains tutorials and guides on how to use DESDEO and also on multiobjective optimization basics in general.

DESDEO is continuously developed by the Multiobjective Optimization Group at University of Jyväskylä (https://optgroup.it.jyu.fi/)