Managing Geological Uncertainty in Critical Mineral Supply Chains: A POMDP Approach with Application to U.S. Lithium Resources

TL;DR

The paper tackles the problem of securing critical minerals, notably lithium, under geological uncertainty by formulating a POMDP that explicitely models information gathering through exploration and belief updates. It demonstrates, via a U.S. lithium supply-chain case, that POMDP-based policies outperform traditional deterministic and stochastic optimization when initial reserve estimates are imperfect, thanks to active uncertainty reduction and adaptive planning. Key innovations include a Gaussian belief representation over site reserves updated with Kalman filtering and the use of online solvers (POMCPOW, DESPOT) to derive scalable, robust sourcing strategies. The work highlights the policy relevance of balancing domestic resource development with international diversification while reducing emissions and satisfying demand, and it sets a foundation for extending the framework to broader minerals and market dynamics.

Abstract

The world is entering an unprecedented period of critical mineral demand, driven by the global transition to renewable energy technologies and electric vehicles. This transition presents unique challenges in mineral resource development, particularly due to geological uncertainty-a key characteristic that traditional supply chain optimization approaches do not adequately address. To tackle this challenge, we propose a novel application of Partially Observable Markov Decision Processes (POMDPs) that optimizes critical mineral sourcing decisions while explicitly accounting for the dynamic nature of geological uncertainty. Through a case study of the U.S. lithium supply chain, we demonstrate that POMDP-based policies achieve superior outcomes compared to traditional approaches, especially when initial reserve estimates are imperfect. Our framework provides quantitative insights for balancing domestic resource development with international supply diversification, offering policymakers a systematic approach to strategic decision-making in critical mineral supply chains.

Figures (8)

• Figure 1: Lithium supply chain example from foreign mining sites to processing plant and manufacturing facilities in the U.S. The lithium mineral is transported by sea from Australia to the U.S. West Coast, then by rail to Nevada. The route covers at least 17,062 nautical miles and takes approximately 10 weeks on average ports_searoute.
• Figure 2: Domestic lithium reserves. The size of the markers represents the estimated volume of reserves at each site. The transparency of the markers represents the uncertainty of the estimates.
• Figure 3: Foreign lithium reserves. The size of the markers represents the estimated volume of reserves at each site. The transparency of the markers represents the uncertainty of the estimates.
• Figure 4: Results of the random policy. It is clear that the policy is not optimized for any of the objectives.
• Figure 5: Results of the greedy policy. The policy is optimized for the amount of LCE processed, and avoiding domestic mining penalties.