Skip to content

Risk Propagation Network + Failure Mode Manifestation

Author: PlanExe Team
Date: 2026-02-10
Status: Proposal
Tags: risk, propagation, failure-modes, simulation, dependencies

Pitch

Model how local risks propagate through dependencies to system-level failure, then simulate manifestation paths across many runs to surface the most likely cascades and highest-leverage interventions.

Why

Risks rarely fail in isolation. Large project failures typically emerge from interacting risks across domains (technical, procurement, financing, regulatory). A propagation model makes these interactions explicit and actionable.

Problem

  • Risk registers treat items independently.
  • Teams under-estimate compounding effects.
  • Mitigation choices are not ranked by systemic impact.

Proposed Solution

Build a Risk Propagation Network that:

  1. Represents risks, tasks, and milestones as a connected graph.
  2. Encodes causal links and delay effects between nodes.
  3. Simulates cascades across the network.
  4. Outputs failure pathways and intervention leverage scores.

Architecture

Plan JSON
  -> Risk + Dependency Extraction
  -> Propagation Graph Builder
  -> Cascade Simulator (Monte Carlo)
  -> Failure Path Analyzer
  -> Mitigation Prioritizer

Graph Model

  • Nodes: risks, tasks, milestones, resources.
  • Edges: causal amplification and delay links.
  • Weights: probability impact, lag time, and severity multiplier.

Example edge

  • Procurement delay -> schedule slippage (weight: high, lag: 2 weeks)
  • Schedule slippage -> financing drawdown risk (weight: medium, lag: 1 month)

Simulation

Run multi-step simulations to reveal cascades:

  • Sample risk events based on probability distributions.
  • Propagate effects through graph edges.
  • Track which nodes fail, when, and how often.

Outputs per run:

  • failure sequence
  • time-to-failure
  • cost impact

Output Schema

{
  "top_failure_paths": [
    {
      "path": ["procurement_delay", "schedule_slip", "financing_gap"],
      "probability": 0.18,
      "expected_loss": 4200000
    }
  ],
  "intervention_points": [
    {"node": "procurement_delay", "leverage": 0.72}
  ]
}

Integration Points

  • Feeds into Monte Carlo plan success probability engine.
  • Adds a propagation-adjusted risk score to plan ranking.
  • Triggers mitigation playbooks for top cascades.

Success Metrics

  • Reduction in surprise compound failures.
  • Increased mitigation effectiveness vs baseline risk registers.
  • Improved forecast accuracy for delays and cost overruns.

Risks

  • Model complexity could obscure interpretation.
  • Missing edges lead to false security.
  • Overfitting to historical cascades.

Future Enhancements

  • Automated edge discovery from historical plans.
  • Dynamic updates as execution data arrives.
  • Cross-project risk propagation benchmarking.

Detailed Implementation Plan

Phase 1: Graph Construction Layer

  1. Define canonical node types:

  2. risk_event, task, milestone, resource_constraint

  3. Define edge semantics:

  4. causal amplification
  5. schedule delay transfer
  6. cost transfer

  7. confidence score per edge

  8. Build graph extraction adapters from plan artifacts:

  9. WBS + dependencies
  10. risk register
  11. finance assumptions

Phase 2: Propagation Simulator

  1. At each simulation tick:
  2. sample active risk events
  3. propagate effects along outgoing edges

  4. update impacted task states and milestone forecasts

  5. Capture cascade traces:

  6. first-trigger node
  7. propagation chain

  8. terminal failure state

  9. Aggregate over 10,000 runs:

  10. pathway frequencies
  11. expected loss per pathway
  12. median time-to-failure

Phase 3: Mitigation Optimizer

  1. Score intervention points by marginal risk reduction.
  2. Recommend top mitigation portfolio under budget constraints.
  3. Re-simulate with mitigations applied to show deltas.

Suggested algorithmic approach

  • Use weighted directed graph with event queue.
  • Compute influence centrality to prioritize mitigation.
  • Run counterfactual analysis: remove/attenuate edge and measure probability delta.

Data model additions

  • risk_graph_nodes (run_id, node_id, node_type, metadata)
  • risk_graph_edges (run_id, src, dst, edge_type, weight, lag)
  • risk_cascade_paths (run_id, path_json, probability, expected_loss)

Integration points

  • Feed risk pathway penalties into ELO/selection ranking.
  • Push high-risk cascade alerts into governance dashboard.
  • Link mitigation actions back into planning artifacts.

Validation checklist

  • Synthetic graph tests with known cascades.
  • Stability tests for edge-weight perturbation.
  • Human review of top-10 pathways for interpretability.

Detailed Implementation Plan (Graph Operations)

Graph Build Pipeline

  1. Extract nodes from tasks, risks, resources.
  2. Infer edges from dependencies and causal templates.
  3. Score edge confidence and impact weight.

Runtime

  • Run cascade simulation alongside baseline Monte Carlo.
  • Persist top failure chains and intervention candidates.

Mitigation Planner

  • Compute marginal risk reduction for each intervention.
  • Recommend portfolio under mitigation budget constraint.

Governance

  • Require review for interventions with high operational disruption.