Project description
Innovative machine learning framework for decision-making
In recent years, investments have faced increasing challenges as investment decisions have become more complex due to the numerous visible and hidden parameters involved. This complexity has had a significant impact on the decision-making process for long-term investments, particularly in the context of green transition decisions. The ERC-funded DECIDE project aims to create an innovative framework for investment decision-making. This framework will leverage machine learning and operations research to address the growing uncertainty. The project will employ an iterative process, using deep generative models to generate and solve new scenarios, providing users with a range of investment options.
Objective
Many important decisions are taken under uncertainty since we do not know the development of various parameters. In particular the ongoing green transition requires large and urgent societal investments in new energy modes, infrastructure and technology. The decisions are spanning over a very long time-horizon, and there are large uncertainty towards energy prices, demand of energy, and production from renewable sources. Such problem can be described as two-stage stochastic optimization problems, where we first decide which facilities to establish, and then we have to schedule the production/transportation for a stochastic demand, using the given facilities. If the decision variables are discrete, such problems are extremely difficult to solve. In this project we will develop a new framework for investment decision making under uncertainty based on a combination of machine learning and operations research. Instead of solving a complex stochastic optimization problem defined on a fixed set of forecasted scenarios, we propose to use an iterative process: We repeatedly generate new scenarios, solve them using advanced optimization methods, and find the corresponding investment solutions. Our novel way of optimization will use deep generative models (DGMs) to generate small sets of scenarios matching the real distribution, and use a guided local search process to select scenarios that properly reflect properties of the full set of scenarios. The outcome of the iterative process is a palette of near-optimal solutions, which can be analyzed using data science methods to extract associations in investments, outrank dominated choices, and organize investments according to urgency. Knowing the full spectrum of possible choices opens up for a much broader discussion of investments, while allowing soft constraints to also be taken into account. This will enable a more transparent and inclusive decision process, while ensuring well-founded and more robust investment decisions.
Fields of science
Keywords
Programme(s)
- HORIZON.1.1 - European Research Council (ERC) Main Programme
Topic(s)
Funding Scheme
HORIZON-ERC - HORIZON ERC GrantsHost institution
2800 Kongens Lyngby
Denmark