neat/evaluate/objectives

Objective-registration helpers for the NEAT evaluate chapter.

This chapter owns the narrow policy that automatically registers an entropy objective during evaluation when multi-objective mode is active and the controller requests auto-entropy behavior.

The boundary is intentionally policy-sized rather than subsystem-sized. Evaluation does not try to own all objective management here. Instead it answers one narrow question: when the controller is already running in multi-objective mode, should this pass ensure that structural entropy is one of the active objectives?

Read this chapter when you want to answer questions such as:

• Why does auto-entropy registration happen during evaluation instead of in constructor setup or the broader objectives chapter?
• What conditions must be true before entropy is injected automatically?
• Why does this helper invalidate the cached objective list after registration?
• Which controller assumptions stay stable for later ranking and selection reads?

The mental model is a three-step policy gate:

1. check whether multi-objective mode and auto-entropy are both active,
2. skip if entropy is already present or dynamic-objective mode owns the policy,
3. register entropy and invalidate cached objective resolution.

This boundary preserves the current scores, novelty evidence, and species state. It only updates the objective vocabulary that later ranking stages may consume.

flowchart TD
  Evaluate[Evaluation pass reaches auto-objective stage] --> Gate[Check multi-objective and auto-entropy flags]
  Gate --> Skip[Skip when disabled or dynamic mode owns objectives]
  Gate --> Existing[Check whether entropy objective already exists]
  Existing --> Register[Register entropy objective]
  Register --> Invalidate[Invalidate cached objective list]
  Invalidate --> Next[Later ranking reads updated objective set]

neat/evaluate/objectives/evaluate.objectives.ts

registerEntropyObjective

registerEntropyObjective(
  controller: NeatControllerForEval,
): void

Register the entropy objective and invalidate the cached objective list.

Registering the objective is only half of the job. The cached objective list must also be invalidated so later ranking stages resolve the updated objective set instead of continuing to use stale ordering information.

Parameters:

• controller - - NEAT controller instance for evaluation.

runAutoEntropyObjectiveInjection

runAutoEntropyObjectiveInjection(
  controller: NeatControllerForEval,
  evaluationOptions: { [key: string]: unknown; fitnessPopulation?: boolean | undefined; clear?: boolean | undefined; novelty?: { enabled?: boolean | undefined; descriptor?: ((genome: GenomeForEvaluation) => number[]) | undefined; k?: number | undefined; blendFactor?: number | undefined; archiveAddThreshold?: number | undefined; } | undefined; entropySharingTuning?: { enabled?: boolean | undefined; targetEntropyVar?: number | undefined; adjustRate?: number | undefined; minSigma?: number | undefined; maxSigma?: number | undefined; } | undefined; entropyCompatTuning?: { enabled?: boolean | undefined; targetEntropy?: number | undefined; deadband?: number | undefined; adjustRate?: number | undefined; minThreshold?: number | undefined; maxThreshold?: number | undefined; } | undefined; autoDistanceCoeffTuning?: { enabled?: boolean | undefined; adjustRate?: number | undefined; minCoeff?: number | undefined; maxCoeff?: number | undefined; } | undefined; multiObjective?: { enabled?: boolean | undefined; autoEntropy?: boolean | undefined; dynamic?: { enabled?: boolean | undefined; } | undefined; } | undefined; speciation?: boolean | undefined; targetSpecies?: number | undefined; compatAdjust?: boolean | undefined; speciesAllocation?: { extendedHistory?: boolean | undefined; } | undefined; sharingSigma?: number | undefined; compatibilityThreshold?: number | undefined; excessCoeff?: number | undefined; disjointCoeff?: number | undefined; },
): void

Inject the entropy objective when multi-objective evaluation requests it.

This is the controller-facing entrypoint for the evaluate-objectives stage. It behaves like best-effort policy maintenance after the rest of the evaluation evidence has been gathered.

The helper preserves several important controller assumptions:

• current scores are left intact,
• novelty and diversity evidence are left intact,
• current species state is left intact,
• only the objective-registration layer and its cache are updated.

Parameters:

• controller - - NEAT controller instance for evaluation.
• evaluationOptions - - Options object for the current evaluation pass.

Example:

runAutoEntropyObjectiveInjection(controller, controller.options);
console.log(controller._getObjectives?.().map((objective) => objective.key));

shouldAutoInjectEntropy

shouldAutoInjectEntropy(
  evaluationOptions: { [key: string]: unknown; fitnessPopulation?: boolean | undefined; clear?: boolean | undefined; novelty?: { enabled?: boolean | undefined; descriptor?: ((genome: GenomeForEvaluation) => number[]) | undefined; k?: number | undefined; blendFactor?: number | undefined; archiveAddThreshold?: number | undefined; } | undefined; entropySharingTuning?: { enabled?: boolean | undefined; targetEntropyVar?: number | undefined; adjustRate?: number | undefined; minSigma?: number | undefined; maxSigma?: number | undefined; } | undefined; entropyCompatTuning?: { enabled?: boolean | undefined; targetEntropy?: number | undefined; deadband?: number | undefined; adjustRate?: number | undefined; minThreshold?: number | undefined; maxThreshold?: number | undefined; } | undefined; autoDistanceCoeffTuning?: { enabled?: boolean | undefined; adjustRate?: number | undefined; minCoeff?: number | undefined; maxCoeff?: number | undefined; } | undefined; multiObjective?: { enabled?: boolean | undefined; autoEntropy?: boolean | undefined; dynamic?: { enabled?: boolean | undefined; } | undefined; } | undefined; speciation?: boolean | undefined; targetSpecies?: number | undefined; compatAdjust?: boolean | undefined; speciesAllocation?: { extendedHistory?: boolean | undefined; } | undefined; sharingSigma?: number | undefined; compatibilityThreshold?: number | undefined; excessCoeff?: number | undefined; disjointCoeff?: number | undefined; },
): boolean

Check whether the evaluation pass should auto-register the entropy objective.

Automatic injection is intentionally conservative. Dynamic-objective mode is treated as the stronger owner when it is enabled, which prevents this helper from silently competing with a more explicit objective-management policy.

Parameters:

• evaluationOptions - - Options object for the current evaluation pass.

Returns: True when entropy should be injected.