Multi-response Event Tree Analysis
Also known as: MR-ETA, multi-output event tree analysis, multi-response ETA, probabilistic event tree with multiple responses
Multi-response Event Tree Analysis (MR-ETA) extends classical event tree analysis by simultaneously tracking multiple system performance or safety response variables across all accident sequences. Instead of evaluating a single outcome (e.g., probability of failure), it propagates several concurrent response metrics — such as damage severity, downtime, cost, and environmental impact — through the event tree branches, enabling richer risk characterization and trade-off decisions under a single probabilistic framework.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use MR-ETA when a system's failure can produce consequences that are important across several independent dimensions (safety, cost, environment, regulatory), and those dimensions must be weighed simultaneously rather than sequentially. It is particularly appropriate in nuclear, aerospace, chemical process, and critical infrastructure engineering where regulators or stakeholders require multi-dimensional risk profiles. Do not use it when only a single consequence metric is relevant (standard ETA suffices), when the event tree has too few branches to justify the overhead of multiple response models, or when consequence data for the additional response variables are unavailable or highly uncertain — in such cases the false precision of added responses may mislead rather than inform.
Strengths & limitations
- Captures multiple simultaneous risk dimensions (safety, cost, downtime, environment) in one coherent probabilistic model.
- Preserves the visual clarity and cause-effect logic of standard event tree analysis while extending its information content.
- Enables multi-criteria prioritization — high-risk sequences can be identified not just by probability but by their combined burden across responses.
- Directly integrates with fault tree analysis for barrier-failure probabilities, maintaining compatibility with established safety engineering workflows.
- Supports regulatory reporting in industries that require multi-attribute consequence assessment (e.g., nuclear, chemical process).
- Data requirements multiply with each additional response variable; poor consequence data for any one dimension can distort the multi-response profile.
- Correlations among response variables (e.g., high economic loss and high downtime tend to co-occur) are difficult to model within a simple branch-probability framework.
- Aggregating across heterogeneous response dimensions (lives, dollars, tonnes of emissions) requires explicit value judgments that may be contested by stakeholders.
- The combinatorial growth of sequences in large trees (many safety functions) makes thorough multi-response evaluation computationally and analytically demanding.
Frequently asked
How is MR-ETA different from standard event tree analysis?
Standard ETA produces one probability or consequence estimate per accident sequence. MR-ETA attaches a vector of response values — safety consequences, cost, downtime, environmental impact — to each sequence. This turns each leaf node from a single number into a multi-dimensional outcome profile, enabling risk ranking by any combination of responses rather than by a single metric.
Do I need a separate model for each response variable?
Yes, in general. Each response variable (e.g., number of injuries, economic loss) requires its own consequence model conditioned on the scenario. The event tree structure and branch probabilities are shared, but the consequence functions are specific to each response dimension. This is the main additional effort compared with standard ETA.
Can I combine MR-ETA with design of experiments (DoE)?
Yes. DoE can be used to efficiently explore how design parameters (barrier reliability, material choices, control system settings) affect the multi-response outcome profile across event tree sequences. Response surface methods or Taguchi arrays applied to the branch probability inputs generate a surrogate model that predicts the full multi-response risk profile as a function of design variables, enabling design optimization under risk constraints.
How do I handle uncertainty in branch probabilities?
Monte Carlo propagation is the most common approach: branch probabilities are represented as probability distributions (log-normal or beta are typical) and sampled repeatedly, producing uncertainty bands on each response variable per sequence. This converts point estimates into confidence intervals and reveals which branches drive output uncertainty most strongly — a sensitivity analysis step that is especially valuable in multi-response settings.
When is fault tree analysis preferable to MR-ETA?
Fault tree analysis (FTA) is preferable when the primary question is how a specific top-level failure can occur — tracing causes from effect to root. ETA (and MR-ETA) is preferable when the question is what consequences follow from an initiating event — tracing effects forward from cause. For complex systems, FTA and MR-ETA are often used together: FTA supplies the barrier-failure probabilities that feed into the event tree branches.
Sources
How to cite this page
ScholarGate. (2026, June 3). Multi-response Event Tree Analysis. ScholarGate. https://scholargate.app/en/experimental-design/multi-response-event-tree-analysis
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Event Tree AnalysisReliability↔ compare
- Failure Mode and Effects AnalysisExperimental design↔ compare
- Fault Tree AnalysisReliability↔ compare
- Multi-response fault tree analysisExperimental design↔ compare
- Statistical Process ControlExperimental design↔ compare