Decision-Focused Forecasting: A Differentiable Multistage Optimisation Architecture

TL;DR

This paper introduces a differentiable multistage optimisation architecture for decision-focused forecasting, enabling better decision-making in sequential, intertemporal problems like energy storage and portfolio management.

Contribution

It presents a novel recurrent differentiable architecture that models multistage decision problems, accounting for intertemporal effects of forecasts, which was not addressed in prior single-stage focused work.

Findings

Outperforms existing approaches in energy storage arbitrage

Achieves superior results in portfolio optimisation tasks

Demonstrates effective gradient adjustments for multistage decision effects

Abstract

Most decision-focused learning work has focused on single stage problems whereas many real-world decision problems are more appropriately modelled using multistage optimisation. In multistage problems contextual information is revealed over time, decisions have to be taken sequentially, and decisions now have an intertemporal effect on future decisions. Decision-focused forecasting is a recurrent differentiable optimisation architecture that expresses a fully differentiable multistage optimisation approach. This architecture enables us to account for the intertemporal decision effects of forecasts. We show what gradient adjustments are made to account for the state-path caused by forecasting. We apply the model to multistage problems in energy storage arbitrage and portfolio optimisation and report that our model outperforms existing approaches.