Supply Chain Disruption Simulation

• Supply Chain Disruption Simulation is a computational modeling approach that integrates SIR-based demand generators with discrete event techniques to capture real-world operational shocks.
• It quantifies the impact of dynamic demand surges and capacity constraints using hybrid architectures, yielding actionable metrics like fulfillment times and cost-resilience tradeoffs.
• The simulation framework guides resilience policies by stress-testing inventory adjustments, transportation mode switching, and regional safety stock strategies under varied disruption scenarios.

A supply chain disruption simulation refers to the computational modeling and analysis of events that interrupt or destabilize the normal flows of materials, information, or capital within supply networks. This domain encompasses a wide spectrum of techniques—ranging from stochastic agent-based models and hybrid system dynamics-discrete event frameworks to optimization-driven simulators and knowledge graph approaches—each aiming to quantify, predict, and manage the operational and strategic impact of disruptions such as pandemics, natural disasters, facility failures, demand surges, or geopolitical events.

1. Hybrid Simulation Architectures for Disruption Modeling

Recent research has established that hybrid simulation architectures are essential for realistically capturing and stress-testing the non-stationary and often systemic nature of disruption phenomena. For instance, in the oxygen concentrator (OC) supply chain, an integrated framework couples an epidemiological SIR (Susceptible-Infected-Recovered) model with a multiscale, discrete event simulation (DES) of the logistics network (Camur et al., 2023). Here, the SIR subsystem generates non-stationary, dynamically-parameterized demand surges based on infection and hospitalization rates, which directly drive demand signals into the DES model responsible for supplier, manufacturer, distributor, and customer interaction (including queuing, (Q, R) inventory control, and routing policies).

This integration enables strict propagation of real-world, event-driven demand shocks (e.g., +78% under aggressive pandemic conditions), manifesting in detailed operational metrics such as order fulfillment times, distributor backlogs, and cost-resilience tradeoffs as policies (dynamic inventory, transportation mode switching, regionally-increased safety stocks) are varied.

2. Modeling Demand Shocks and Propagation Dynamics

Effective simulation of supply chain disruptions requires explicit mathematical treatment of dynamic, sometimes endogenous, demand signals. Augmented SIR models can incorporate time-varying infection rates, intervention effects, and reinfection (waning immunity), and may model hospitalization compartments and hospital-level replenishment triggers. The governing ODEs typically follow the structure:

$$\begin{align*} \frac{dS}{dt} &= -\beta S I / N \ \frac{dI}{dt} &= \beta S I / N - \gamma I \ \frac{dR}{dt} &= \gamma I \end{align*}$$

with extensions for hospitalization, death, and immunity decline cycles. The generated demand (e.g., OCs per time unit) feeds directly into discrete event supply chain modules, enabling high-fidelity, non-stationary stress testing (Camur et al., 2023).

In the absence of demand-driven disruptions, propagation can be simulated along a directed, weighted network (Leontief or knowledge graph models), with each node updating state according to supply, demand, or operational dependencies. Models may include probabilistic failure propagation, e.g., via

$$\rho_v(t) = 1 - \prod_{q \in \bigcup_{i\in U_v}P^{\text{out}_i}} (1 - p_{vqt})$$

where $\rho_v(t)$ is the probability a node $v$ fails due to the set $U_v$ of newly inactivated suppliers, each potentially propagating the shock through specific products $q$ (Inoue et al., 8 Oct 2024).

3. Disruption Scenarios, Resilience Policies, and Operational Strategies

Simulation frameworks are typically evaluated under a suite of disruption scenarios:

• Demand shocks: sudden, SIR-driven surges in hospital equipment needs (Camur et al., 2023).
• Capacity disruptions: modeled as reductions in manufacturing, storage, or transportation capacity due to localized or systemic shocks.
• Lead time perturbations: increases in supplier lead times, which can trigger system-wide lateness and service level degradations (Estrada-Garcia et al., 2023).

Resilience policies tested include:

• Static vs. dynamic inventory policies: Static (pre-disruption, stationary) policies fail under non-stationary demand; dynamically adjusting (Q, R) parameters in distributors/hospitals ensures agile backorder reduction and leads to superior resilience metrics under shock (Camur et al., 2023).
• Transportation mode switching: Integration of expedited (e.g., air freight) options for replenishment, particularly for long-distance links, can halve (or better) high-percentile fulfillment times during surges, but at increased cost.
• Regional safety stock adjustment: Small or sparsely-covered regions are particularly vulnerable to stockouts under synchronous surges; localized increases in safety stock introduce negligible system-wide costs but sharply improve fulfillment times in tail-risk scenarios.
• Backward and forward recovery strategies: Adopting surplus inventory, dynamic formation of new supplier-client links, or backup supplier arrangements can prevent or soften phase transitions in cascade models by rapidly re-routing flows after failures (Yang et al., 2019).

4. Numerical Results and Simulation Outcomes

Simulation frameworks supporting non-stationary, event-coupled demand reveal several key results:

Scenario / Policy	Fulfillment Time	% Demand Met	Cost Impact	Vulnerable Regions
Static (pre-COVID) inventory	Several weeks	<80%	Baseline	Small/low-stock states
Dynamic (demand-tracking)	Days	95–99%	+inventory, +transp.	Improved, still fragile
Air freight (long-haul)	~2× reduction	>95%	+transp.	Mitigates backlog
Increased safety stock (small)	<3 days	>98%	Negligible	Strong improvement

Source: Factual statements from (Camur et al., 2023), Table 2 and simulation results

These findings establish that static preparedness policies are inadequate for pandemic-scale, non-stationary demand shocks, and that the adoption of dynamically-adaptive, geographically-targeted resilience strategies provides a measurable reduction in fulfillment delays and backlog accumulation.

Moreover, analysis of simulation outputs highlights system vulnerabilities: in highly-populated states, large, early orders can deplete manufacturer stocks and trigger cascading shortages in smaller regions ("first-in-queue" effect), warranting dynamic allocation and prioritization policies.

5. System Architecture and Computational Implementation

The implementation of an integrated disruption simulation system comprises two tightly coupled modules:

• SIR Epidemiological Demand Generator: Parameterized at the region (state) level, incorporating dynamic rates and stochastic hospitalizations. Demand for critical medical devices is calculated as a function of evolving hospitalization statistics. Stochastic agent-based enhancements allow for non-exponential (i.e., realistic) distributions of hospital stays and resource use.
• Discrete Event Supply Chain Simulator: Models entities (manufacturers, distributors, hospitals) with detailed inventory, order, and replenishment protocols; incorporates process delay distributions, capacity constraints, and transport logistics (including air/ground mode switching). Network "illumination" via supplier mapping produces realistic supply flows across 278+ identified suppliers and 52 regional distribution centers (Camur et al., 2023).

These modules are synchronized via a demand signal interface: the SIR model computes expected and realized order arrivals, which immediately informs the queueing and inventory system in the DES. Policy scenarios can be batch-simulated for rapid high-dimensional sensitivity analysis.

6. Implications and Applications

Hybrid, event-driven simulation of supply chain disruptions provides a rigorous foundation for policy stress testing, operational optimization, and strategic preparedness. For critical equipment supply chains, such as medical devices, this approach enables:

• Performance-based comparison of alternative resilience interventions (inventory, logistics, safety stock) under scenario uncertainty
• Rapid identification of high-risk bottlenecks and dynamic vulnerabilities (e.g., cascading shortages, spatial inequity)
• Quantification of trade-offs between service levels, response time, and total cost for resilient system design

This methodological paradigm has direct utility for health system preparedness, infrastructure, and critical supply chain management, as well as for the real-time operational planning of distributed enterprises during emergent disruptions.

7. Limitations and Directions for Further Research

While integrated SIR-DES simulation architectures advance the state of disruption modeling, several challenges remain:

• Precision of parameterization depends on data quality for both epidemiological and supply chain modules.
• Agent-based extensions for more granular modeling (e.g., hospital-level choice, consumer hoarding) can enrich behavioral realism but are computationally intensive.
• Incorporating cross-region feedbacks and supply chain competition requires further development.
• Extension to multi-product, multi-resource settings and full inter-supply chain effects (e.g., critical component shortages propagating indirectly) is needed for generalized robustness analysis.

A plausible implication is that ongoing methodological integration with knowledge graph-based simulation, scenario-based benchmarking, and human-in-the-loop analytics will increase both realism and operational relevance for complex, disruption-prone supply networks.

---

Key reference: An Integrated System Dynamics and Discrete Event Supply Chain Simulation Framework for Supply Chain Resilience with Non-Stationary Pandemic Demand (Camur et al., 2023)