Agent-based modeling for by-product metal supply—a case study on indium

Abstract

With rapid development and deployment of clean energy technology, demand for certain minor metals has increased significantly. However, many such metals are by-products of various host metals and are economically infeasible to extract independently. Meanwhile, by-product metals present in the mined ores may not be extracted even if they are sent to smelters along with host metal concentrates if it is not economically favorable for the producers. This dependency poses potential supply risks to by-product metals. Indium is a typical by-product metal, mainly from zinc mining and refining, and is important for flat panel displays, high efficiency lighting, and emerging thin-film solar panel production. Current indium supply–demand forecast models tend to overlook the volatile and competitive nature of minor metal market and are mostly based on top-down ap-proaches. Therefore, a bottom-up agent-based model can shed new light on the market dynamics and possible outcome of future indium supply–demand relationship. A multi-layered model would also be helpful for identifying possible bottlenecks of indium supply and finding solutions. This work takes indium as an example of minor metal market and sets up an agent-based model to pre-dict future market situation and supply–demand balance. The market is modeled as a Cournot com-petition oligopolistic market by refineries with capacity restriction based on host metal production. The model maintains active Nash equilibrium each year to simulate competitions between suppli-ers. The model is validated and verified by historical data and sensitivity analysis. Several scenarios are also explored to illustrate possible uncertainties of the market.