Introduction
DESDEO
Decision Support for computationally Demanding Optimization problems
DESDEO is a free and open-source Python-based framework for developing and experimenting with interactive multiobjective optimization.
DESDEO contains implementations of some interactive methods and modules that can be utilized to implement further methods.
We welcome you to utilize DESDEO and develop it further with us.
DESDEO brings interactive methods closer to researchers and practitioners worldwide by providing them with implementations of interactive methods.
DESDEO is part of DEMO (Decision analytics utilizing causal models and multiobjective optimization which is the thematic research area of the University of Jyväskylä.
Mission
The mission of DESDEO is to increase awareness of the benefits of interactive methods make interactive methods more easily available and applicable. Thanks to the open architecture, interactive methods are easier to be utilized and further developed. The framework consists of reusable components that can be utilized for implementing new methods or modifying the existing methods. The framework is released under a permissive open source license.
Multiobjective optimization
In multiobjective optimization, several conflicting objective functions are to be optimized simultaneously. Because of the conflicting nature of the objectives, it is not possible to obtain individual optima of the objectives simultaneously but one must trade-off between the objectives.
Interactive methods in multiobjective optimization
Interactive methods are iterative by nature where a decision maker (who has substance knowledge) can direct the solution process with one's preference information to find the most preferred balance between the objectives. In interactive methods, the amount of information to be considered at a time is limited and, thus, the cognitive load set on the decision maker is not too demanding. Furthermore, the decision maker learns about the interdependencies among the objectives and also the feasibility of one’s preferences.
The Research Projects Behind DESDEO
About the Multiobjective Optimization Group
The Multiobjective Optimization Group developes theory, methodology and open-source computer implementations for solving real-world decision-making problems. Most of the research concentrates on multiobjective optimization (MO) in which multiple conflicting objectives are optimized simultaneously and a decision maker (DM) is supported in finding a preferred compromise.
About the DESDEO research project
DESDEO contains implementations of some interactive methods and modules that can be utilized to implement further methods. DESDEO brings interactive methods closer to researchers and practitioners world-wide, by providing them with implementations of interactive methods.
Interactive methods are useful tools for decision support in finding the most preferred balance among conflicting objectives. They support the decision maker in gaining insight in the trade-offs among the conflicting objectives. The decision maker can also conveniently learn about the feasibility of one’s preferences and update them, if needed.
DESDEO is part of DEMO (Decision analytics utilizing causal models and multiobjective optimization) which is the thematic research area of the University of Jyväskylä (jyu.fi/demo).
We welcome you to utilize DESDEO and develop it further with us.
About DAEMON
The mission of DAEMON is method and software development for making better data-driven decisions. The project considers data and decision problems from selected fields as cases to inspire the research and demonstrate the added value.
In DAEMON, we support optimizing conflicting objectives simultaneously by applying interactive multiobjective optimization methods, where a decision maker (DM) incorporates one’s domain expertise and preferences in the solution process. Overall, we model and formulate optimization problems based on data, so that DMs can identify effective strategies to better understand trade-offs and balance between conflicting objectives. In this, we incorporate machine learning tools, visualize trade-offs to DMs and consider uncertainties affecting the decisions.