State-Augmented Graphs for Circular Economy Triage

TL;DR

The paper addresses end-of-life product triage in the circular economy by extending disassembly sequencing planning with a state-augmented graph that preserves the Markov property, enabling exact dynamic programming or rollout methods for adaptive decision-making. It defines a condition-aware utility combining revenue and processing cost under feasibility constraints, and demonstrates the approach on EV battery hierarchies to reveal how history gates CE routing decisions. The worked EV battery examples show how pack- vs module-level decisions depend on health diagnostics and costs, exposing economic-environmental misalignments that policy could target. The framework is designed to be generalizable across products and contexts and points toward future work in uncertainty, reinforcement learning, human-in-the-loop guidance, and multi-agent coordination across facilities.

Abstract

Circular economy (CE) triage is the assessment of products to determine which sustainable pathway they can follow once they reach the end of their usefulness as they are currently being used. Effective CE triage requires adaptive decisions that balance retained value against the costs and constraints of processing and labour. This paper presents a novel decision-making framework as a simple deterministic solver over a state-augmented Disassembly Sequencing Planning (DSP) graph. By encoding the disassembly history into the state, our framework enforces the Markov property, enabling optimal, recursive evaluation by ensuring each decision only depends on the previous state. The triage decision involves choices between continuing disassembly or committing to a CE option. The model integrates condition-aware utility based on diagnostic health scores and complex operational constraints. We demonstrate the framework's flexibility with a worked example: the hierarchical triage of electric vehicle (EV) batteries, where decisions are driven by the recursive valuation of components. The example illustrates how a unified formalism enables the accommodation of varying mechanical complexity, safety requirements, and economic drivers. This unified formalism therefore provides a tractable and generalisable foundation for optimising CE triage decisions across diverse products and operational contexts.

Figures (1)

• Figure 1: $e_{CE}$ is a possible point at which a CE option could be taken. Blue CE edges are where a decision can be made, and orange is where the node information is indistinct. In reality, all nodes are able to exit disassembly and take a CE route, but for clarity and conciseness, only terminal nodes are shown here. (a) A standard DSP graph $G$, where multiple histories (e.g., $e_0 \rightarrow e_3 \rightarrow e_5$ and $e_0 \rightarrow e_1 \rightarrow e_4 \rightarrow e_5$) converge on the same physical node ($v_5$). For node $v_5$ the choice is ambiguous as the processes already performed are not captured. (b) The proposed state-augmented graph $\tilde{G}$, which 'unrolls' the paths. The states $\tilde{v}_6$ and $\tilde{v}_7$ are distinct, as their history $\tau$ differs, but they still represent the same physical node $v_5$. This enables a decision to be made, allowing the admissibility function to correctly approve a CE action (e.g., 'Reuse') for one state while rejecting it for another.