Forecast & Flex: A Double Safeguard Framework for Production Planning

Abstract

Problem definition: The surge in electric vehicle (EV) sales is fueling a substantial rise in demand for EV batteries, posing a formidable challenge for EV battery manufacturers to effectively manage production planning amidst volatile customer demands and increasing customized product requirements. This challenge also exists across various sectors, including automotive, consumer electronics, and industrial equipment. 

Methodology/Results: To address this challenge, we introduce Forecast & Flex, a double safeguard framework that integrates demand forecasting with a dual-mode production strategy. First, we improve demand forecasting accuracy by combining customers’ forecasts with ensemble learning of state-of-the-art prediction algorithms. This serves as the first layer of safeguard. Second, we hedge the remaining forecast risk by incorporating the forecasted demands into a dual-mode production model that optimally splits total production volume into make-to-stock (MTS) and make-to-order (MTO) quantities. This strategic balance enables manufacturers to reliably meet base demand via MTS and responsively meet unexpected demand via MTO, constituting the second layer of safeguard.

Managerial implications: We implemented the double safeguard framework in the production planning of 38 distinct battery modules for a leading EV battery manufacturer with significant benefits. The deployment increased demand forecasting accuracy to an average level of 94.3% and reduced excess inventory by 19.6%, leading to annual cost savings of US$17 million. Furthermore, it decreased battery production variability by 40%, significantly stabilizing production flows. These achievements highlight the tangible benefits of applying our research in a real-world context, offering a compelling case for broader industry adoption.