Solve bayesian influence diagram

What it's for

Extends Gitrevio from forecasts and isolated optimizers into inspectable executive decision models that connect uncertain operating conditions, action consequences, business utility, current evidence, and the value of learning more.

What you give it

Inputs split into evidence read from your connected systems, calibration your team owns, and numerical controls that affect precision but never the result's meaning.

{
  "actions": [
    {
      "id": "scale"
    },
    {
      "id": "hold"
    }
  ],
  "chance_nodes": [
    {
      "conditional_probabilities": [
        {
          "parent_states": {},
          "probabilities": {
            "high": 0.5,
            "low": 0.5
          }
        }
      ],
      "id": "demand",
      "parents": [],
      "states": [
        "high",
        "low"
      ]
    },
    {
      "conditional_probabilities": [
        {
          "action_id": "scale",
          "parent_states": {
            "demand": "high"
          },
          "probabilities": {
            "failure": 0.09999999999999998,
            "success": 0.9
          }
        },
        {
          "action_id": "scale",
          "parent_states": {
            "demand": "low"
          },

What you get back

This is the actual output of running the example above — computed by the same function the platform calls, not an illustration.

{
  "actions": [
    {
      "action_id": "scale",
      "evidence_probability": 1,
      "expected_utility": 37,
      "regret_to_best": 0
    },
    {
      "action_id": "hold",
      "evidence_probability": 1,
      "expected_utility": 32,
      "regret_to_best": 5
    }
  ],
  "assumptions": [
    "The chance-node DAG, conditional probability tables, additive utility tables, and evidence timing encode the accountable decision owner's current model.",
    "Evidence is observed before action; conditioning on an action consequence is rejected.",
    "Expected value of perfect information is an upper bound before measurement cost, delay, partial accuracy, and implementation constraints."
  ],
  "configuration": {
    "minimum_utility_advantage": 0
  },
  "decision": "action_identified",
  "diagram": {
    "action_dependent_nodes": [
      "outcome"
    ],
    "chance_nodes": 2,
    "joint_worlds": 4,
    "nodes_affected_by_action": [
      "outcome"
    ],
    "observed_evidence": {},
    "utility_nodes": 1
  },
  "executive_summary": {
    "maximum_expected_value_of_perfect_information": 21,
    "most_valuable_unobserved_node": "demand",
    "selected_expected_utility": 37,
    "utility_advantage_over_second": 5
  },
  "information_candidate_count": 1,
  "interpretation": "The selected action maximizes expected governed utility inside the declared influence diagram. A different causal structure, probability calibration, utility scale, or evidence set can change the decision.",

How it works

Sequential Bayesian & bandits — Learn while deciding — update beliefs as evidence arrives and choose where the next unit of effort is worth spending.

1. 1 Validate and topologically order the discrete chance DAG, requiring complete normalized conditional-probability tables for every parent-state and action combination.
2. 2 Enumerate the bounded joint state space, condition only on evidence observed before action, and calculate each action's posterior expected additive utility.
3. 3 Select the action with highest governed expected utility only when its advantage clears the practical decision threshold; retain regret to the best action for every alternative.
4. 4 For each unobserved action-independent node, compute the optimal contingent action by state and its expected value of perfect information as an upper bound on a possible measurement program.

Before you trust it

Every tool in the catalog ships with the conditions under which its answer is meaningful — and the conditions under which it should abstain instead of guessing.

Calibrating it to your org

Calibration workflow

1. 1 Define the management decision, target outcome, aggregate unit, privacy boundary, cadence, and prediction/intervention horizon for this organization.
2. 2 Build a tenant-scoped historical cohort using only information available before each prediction or decision; preserve zero periods, censoring, assignment probabilities, and unresolved outcomes when the method requires them.
3. 3 Estimate statistical parameters on training history, but obtain costs, utilities, risk tolerance, practical-effect thresholds, capacity, and policy constraints from accountable decision owners.
4. 4 Validate on later time windows or held-out aggregate units at the deployment grain, against a simple baseline and the function-specific validation strategy.
5. 5 Deploy only if the returned decision clears evidence, overlap, calibration, robustness, and guardrail checks; warning, unsupported, schema-gap, and fallback decisions are abstentions.
6. 6 Monitor realized outcomes, data drift, coverage, and decision regret; recalibrate at a governed cadence or after a detected regime/definition change, never merely because a stakeholder dislikes the result.

Call it from your AI

You don't wire up 388 tools in your MCP client. The GitRevio MCP server exposes 18 tools, three of which let an agent search the catalog, read a tool's schema, and run it — so the assistant finds this one on its own.

gitrevio_capabilities_search
  { "q": "solve an exact discrete bayesian influence" }
  → finds "solve_bayesian_influence_diagram"

gitrevio_capability_describe
  { "capability_id": "solve_bayesian_influence_diagram" }
  → returns the input schema and agent guidance shown on this page

gitrevio_capability_run
  { "capability_id": "solve_bayesian_influence_diagram", "arguments": { ... } }
  → returns the result shown above

Works in Claude Desktop, Claude Code, Cursor, Cline, Continue.dev, Goose and Aider. See the MCP server.

Related tools

See every tool in Analytics assurance & orchestration →