1 cumulative citations
View corpus contextDecentralized factories are more resilient: self-organized production systems with crisis-aware AI improve absolute resilience by about 10.7% and sustain substantially higher throughput during supply shocks than hierarchical control. Adaptive coalition formation, proactive resource conservation, and preserved coordination structures drive faster recovery and stronger baseline performance.
Modeling Organizational Resilience in Human-Cyber-Physical Systems (Industry 5.0) Through Collective Dynamics, Decision Scenarios and Crisis-Aware AI: A Multi-Method Simulation Approach
openalex
theoretical
low evidence
7/10 relevance
Summary only
summary available; pdf_status=paywall
DOI
Source
PDF
In simulation, self-organized manufacturing systems augmented with crisis-aware Q-learning agents deliver higher baseline throughput, ~10.7% greater absolute resilience, and faster recovery under supply disruptions than centralized or distributed hierarchical controls.
Citation observations
Cumulative provider counts captured on specific dates; providers are never combined.
0 cumulative citations
View corpus contextSupply chain disruptions during the COVID-19 pandemic exposed structural vulnerabilities of centrally controlled manufacturing systems, motivating renewed interest in organizational resilience within the context of Industry 5.0 human–cyber–physical systems. This study investigates how organizational decision-making paradigms and crisis-aware artificial intelligence (AI) jointly influence performance, crisis response, and recovery. An agent-based modeling (ABM) framework is developed to compare centralized, distributed, and self-organized organizational structures across 650 simulation runs under a controlled supply side disruption. A crisis-aware Q-learning architecture enables AI agents to shift from efficiency-oriented to stability-oriented strategies when resource scarcity is detected. To avoid baseline-dependent bias, resilience is evaluated using an absolute, capacity-normalized metric. Results indicate that self-organized systems consistently outperform centralized and distributed structures in baseline performance, crisis throughput, and recovery speed. The integration of crisis-aware AI further increases absolute resilience by approximately 10.7% and enables substantially higher throughput during disruption compared to hierarchical control. Enhanced performance is primarily driven by adaptive coalition formation, proactive resource conservation, and rapid post-crisis recovery supported by preserved coordination structures. These findings provide quantitative support for Industry 5.0’s human-centric principles and show that decentralized decision-making augmented by context-adaptive AI offers a robust organizational design strategy for volatile manufacturing environments.
Summary
Main Finding
Self-organized manufacturing systems augmented with a crisis-aware Q-learning AI substantially outperform centralized and distributed (hierarchical) organizational structures under supply-side disruption. Across simulations, decentralized, self-organizing decision-making plus context-adaptive AI raised absolute, capacity-normalized resilience by ≈10.7%, delivering higher baseline throughput, substantially greater crisis throughput, and faster recovery.
Key Points
- Organizational comparison: three paradigms evaluated — centralized (hierarchical control), distributed, and self-organized (decentralized) systems.
- AI architecture: a crisis-aware Q-learning agent that switches from efficiency-oriented policies to stability-oriented policies when resource scarcity signals are detected.
- Performance gains: self-organized systems led in baseline performance, crisis throughput, and recovery speed. Adding crisis-aware AI increased absolute resilience by ~10.7% versus non-adaptive baselines.
- Mechanisms driving gains:
- Adaptive coalition formation among local agents to re-route and share scarce resources.
- Proactive resource conservation under detected scarcity (stability-oriented policies).
- Rapid post-crisis recovery due to preserved local coordination structures and quicker re-establishment of flows.
- Resilience measurement: used an absolute, capacity-normalized resilience metric to avoid biases tied to baseline performance differences.
- Robustness: findings are robust across 650 simulation runs under a controlled supply-side disruption scenario.
Data & Methods
- Modeling approach: agent-based model (ABM) representing manufacturing/coordination nodes and interactions under Industry 5.0 human–cyber–physical system assumptions.
- Scenarios: controlled supply-side disruption applied across runs; organizational structure parametrized as centralized, distributed, or self-organized.
- AI intervention: crisis-aware Q-learning implemented for agents (training/runtime details not specified in summary) that detects resource scarcity and shifts reward orientation from efficiency to stability.
- Experimental design: 650 simulation runs covering the different organizational paradigms and AI/no-AI configurations to statistically compare baseline performance, throughput during disruption, and recovery trajectories.
- Evaluation metric: absolute, capacity-normalized resilience (so results are not driven by baseline capacity differences); additional metrics include crisis throughput and recovery speed.
Implications for AI Economics
- Organizational design and investment:
- Firms operating in volatile environments should consider decentralized, self-organizing structures augmented with context-adaptive AI as a resilience-enhancing strategy.
- Capital allocation decisions should weigh short-term efficiency gains from centralized control against longer-term resilience dividends from decentralization plus adaptive AI.
- Technology policy and standards:
- Policies that encourage interoperability and local coordination protocols (so local agents can form coalitions) can magnify the resilience benefits of decentralized AI.
- Standards for crisis-aware AI behavior and testing could accelerate safe deployment in critical supply chains.
- Labor and human-centered design:
- Industry 5.0 principles—combining human oversight with cyber-physical autonomy—are supported quantitatively; investments in training and human–AI interfaces remain important to realize benefits.
- Decentralization may shift skill demand toward local coordination, negotiation, and oversight roles.
- Market-level effects and risk management:
- Widespread adoption of decentralized, crisis-aware AI could reduce systemic supply-chain fragility, altering insurance pricing, procurement strategies, and supplier selection criteria.
- Regulators and firms should consider second-order effects (e.g., coordination externalities, strategic withholding) and potential new forms of systemic risk from emergent decentralized interactions.
- Research and evaluation needs:
- Empirical validation: translate ABM insights to real-world pilot studies and field experiments across sectors and disruption types.
- Economic modeling: incorporate resilience gains into cost–benefit frameworks for firm organization, investment, and policy incentives.
- Broader metrics: extend evaluation beyond throughput and recovery to include costs, welfare impacts, labor outcomes, and cross-firm externalities.
Suggested next steps for researchers and practitioners: replicate with diverse disruption types and network topologies, quantify monetary value of the ~10.7% resilience uplift, and test human–AI decision protocols in mixed human–agent environments.
Assessment
Paper Typetheoretical
Evidence Strengthlow — Findings are derived entirely from simulation experiments rather than observational or experimental real-world data; results depend on model specification, parameter choices, agent behavioral rules, and a single disruption scenario, limiting external validity and empirical support.
Methods Rigormedium — The study uses a reasonably rigorous ABM setup (large number of runs, controlled counterfactuals, capacity-normalized resilience metric, and a crisis-aware RL architecture), but lacks reported real-world calibration/validation, details on sensitivity analyses and robustness to alternative parameterizations, and limited modeling of human decision-making or multiple shock types.
SampleSimulated manufacturing ecosystems represented by agent-based models comparing centralized, distributed, and self-organized organizational structures across 650 runs; agents include production units and AI decision agents using a crisis-aware Q-learning policy that switches from efficiency- to stability-oriented actions under resource scarcity; a controlled supply-side disruption is imposed and resilience measured via an absolute, capacity-normalized throughput metric.
Themesorg_design human_ai_collab productivity
IdentificationControlled agent-based model experiments that compare counterfactual organizational architectures (centralized, distributed, self-organized) under a standardized supply-side disruption; causal statements derive from within-model manipulations of structure and the presence/absence of a crisis-aware Q-learning agent policy across 650 simulation runs and normalized resilience metrics.
GeneralizabilityResults are model-dependent and may not hold outside the specific ABM parameterization and agent rules used, Stylized single disruption scenario (supply-side shock) may not represent other shock types (demand shocks, cascading failures), Human factors, institutional constraints, and real-world coordination costs are likely under-modeled, Scaling from simulated units to real manufacturing firms and supply networks may introduce additional frictions, Assumes availability and correct functioning of crisis-aware AI and communication infrastructure that may be cost- or data-constrained in practice
Claims (7)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| An agent-based modeling (ABM) framework was developed to compare centralized, distributed, and self-organized organizational structures across 650 simulation runs under a controlled supply-side disruption. Other | null_result | comparison of organizational structures under disruption (methodological) |
Reading fidelity
high
Study strength
medium
|
n=650
|
| Self-organized systems consistently outperform centralized and distributed structures in baseline performance, crisis throughput, and recovery speed. Organizational Efficiency | positive | organizational performance (baseline performance, crisis throughput, recovery speed) |
Reading fidelity
high
Study strength
medium
|
n=650
|
| A crisis-aware Q-learning architecture enables AI agents to shift from efficiency-oriented to stability-oriented strategies when resource scarcity is detected. Decision Quality | positive | strategy orientation of AI agents (efficiency vs stability) |
Reading fidelity
high
Study strength
medium
|
n=650
|
| Integration of crisis-aware AI increases absolute resilience by approximately 10.7% and enables substantially higher throughput during disruption compared to hierarchical control. Organizational Efficiency | positive | absolute resilience (capacity-normalized) and throughput during disruption |
Reading fidelity
high
Study strength
medium
|
n=650
approximately 10.7% increase
|
| Enhanced performance is primarily driven by adaptive coalition formation, proactive resource conservation, and rapid post-crisis recovery supported by preserved coordination structures. Team Performance | positive | mechanisms driving improved performance (coalition formation, resource conservation, recovery speed) |
Reading fidelity
medium
Study strength
medium
|
n=650
|
| Resilience is evaluated using an absolute, capacity-normalized metric to avoid baseline-dependent bias. Other | null_result | resilience measurement approach |
Reading fidelity
high
Study strength
medium
|
not reported
|
| These findings provide quantitative support for Industry 5.0’s human-centric principles and show that decentralized decision-making augmented by context-adaptive AI offers a robust organizational design strategy for volatile manufacturing environments. Organizational Efficiency | positive | organizational robustness / suitability of design strategy |
Reading fidelity
medium
Study strength
medium
|
n=650
|