Supply-Chain / Network ABM Model
How does a localized supply disruption propagate through a production network to become an aggregate output loss, and what determines whether the shock is absorbed locally or amplified into a systemic bottleneck?
Graph workspace
Multi-panel time-series with network heatmap sidebar
Supply-chain ABMs produce stochastic time series (aggregate output, price level, inventory utilization) alongside spatial network states (which nodes are disrupted, which edges are carrying reduced flow). The canonical visualization pairs a time-series panel showing the output decline and recovery with a heatmap or color-coded adjacency matrix showing disruption propagation through the network. This makes the connection between network topology and aggregate dynamics visible in one view.
Current results
Output Index
Real aggregate output in the final week of the run, relative to the pre-shock baseline: final-goods quantities valued at fixed baseline prices, so a price spike cannot move this index. Values below 100 mean supply is still disrupted at the end; the dip during the disruption itself shows on the chart. Because output is hard-capped at capacity, the index cannot exceed 100.
100.00
index
Disrupted Firms
Fraction of firms producing below 80% of capacity in the final week of the run. After the network recovers it reads 0; the breadth of the cascade while the shock lasts shows on the chart.
0.00%
percent
Price Index
Weighted average price across all firms in the final week of the run, relative to baseline. The cost-push spike during the disruption shows on the chart; by the final week prices have mostly settled.
100.03
index
Avg Inventory Ratio
Average inventory as a fraction of the BASELINE target stock (not the current, post-stockout target), averaged over every week of the run. Below 1.0 indicates economy-wide inventory depletion. It can exceed 1.0 during the restocking wave because a stockout ratchets the target up by a factor each period, bounded at 2x baseline - measured against the current target the ratio could never exceed 1. The page reports the run average rather than the final week because this sketch's inventory line carries an illustrative cycle: over the whole run that cycle cancels, so the readout moves only with the settings and reads exactly 1.0 at the defaults.
1.23
ratio
Controls
Shock severity: 0.8
The share of production capacity lost at a disrupted firm, up to a complete shutdown.
Higher shock fraction produces deeper output trough and longer recovery. At 1.0, downstream firms fully deplete inventory within buffer-stock weeks.
Inventory buffer (weeks): 4
How many weeks of supplies each firm holds; bigger stockpiles delay the disruption spreading.
With zero inventory, the output drop is immediate. With 8+ weeks, the cascade is delayed and dampened, giving time for the disrupted node to recover before downstream firms are affected.
Substitution elasticity: 0.2
How easily a firm can reroute to another supplier; at zero, no rerouting is possible.
At , the full cascade propagates. As rises toward 1.0, firms partially substitute around disrupted suppliers and the output trough is shallower.
Price adjustment speed: 0.1
How fast firm prices respond when demand runs ahead of what they can supply.
A higher produces a sharper price spike at disrupted nodes that propagates downstream as cost-push inflation. The spike shows on the chart; the readouts are final-week values, taken after prices have settled, so they barely move.
Network density: 0.15
How many of the possible supplier-customer links exist; more links mean firms spread sourcing wider.
Denser networks diffuse disruptions across more paths, reducing the depth but broadening the reach of the cascade. Very sparse networks concentrate impact on a few downstream chains.
Shock breadth (fraction of firms hit): 0.03
The share of firms hit; at usual settings the blow lands on the middle of the chain.
Raising breadth converts a localized cascade into a systemic one: inventory buffers cannot absorb a shock that hits suppliers and customers at the same time, so the output trough deepens and the recovery lengthens even at unchanged severity.
The margin firms add over cost, which sets how much of a supplier price rise they pass on.
A higher raises per-hop pass-through, so an upstream price spike survives more tiers before it dies out. Because averages 0.25, even leaves three-hop pass-through at about 5% - the visible price-index moves come mostly from the scarcity term, not from cost pass-through.
Order safety margin: 0.25
The extra cushion firms add to their stock target; with none, orders simply match usage.
A higher amplifies order swings relative to demand swings. A 10% rise in the demand forecast lifts one period's order by roughly - about 60% at the 4-week default and - and the amplification compounds tier by tier upstream.