{
  "handle": "Optimize",
  "mechanism": "The iterative process of adjusting parameters or structure to maximize (or minimize) a specific Objective Function defined by a {{metric}}. It generates candidate {{solution}}s, applies the definition to obtain comparable {{metric_reading}} values, and retains the best candidate. It can be Local ({{gradient}} Descent) or {{global}} (Evolutionary).",
  "gloss": "Maximize utility against a metric",
  "failure_modes": [
    "Overfitting: {{solution}} works perfectly on test data but fails in reality.",
    "Goodhart's Law: Optimizing for a proxy {{metric}} destroys the actual value.",
    "Local Optima: Getting stuck in a 'good enough' state that prevents finding the best state.",
    "Premature Optimization: Optimizing before the problem is fully understood."
  ],
  "invariants": [
    "Monotonicity: no step degrades the best-known {{solution}}. A step that finds nothing better retains the incumbent.",
    "Measurability: the Objective Function must be quantifiable via a {{metric}}."
  ],
  "preconditions": [
    "Objective Function is defined",
    "Candidate space is explorable",
    "Baseline {{metric_reading}} is known"
  ],
  "postconditions": [
    "Best Candidate {{solution}} found",
    "Best observed {{metric_reading}} is at least as good as the Baseline reading"
  ],
  "_meta": {
    "tier": 1,
    "related": [
      "EvaluatorOptimizer",
      "RegretMinimization",
      "ParetoFront"
    ],
    "ring": 1,
    "supersedes": [
      "sema:Optimize#mh:SHA-256:30759374ede0314268d795a5a9878ac68876d0695c4ce0847e615da2b9fedf32",
      "sema:Optimize#mh:SHA-256:94e09b6b86c58500e9716efa1c955fc115e5179b95e48bd45c980af77d7c4c17",
      "sema:Optimize#mh:SHA-256:8a1d872b1e5019a0f4ef856d5f6c6809b87271838a50e55ef2ec21835839b7eb"
    ],
    "path": [
      "Mind",
      "Strategy"
    ]
  },
  "sema_id": "sema:Optimize#mh:SHA-256:38771f5e8f583966cfb8af8ff0d2e6737c8d0bf87329ddc49d8049450e852674",
  "sema_ref": "Optimize#3877",
  "sema_stub": "3877",
  "dependencies": {
    "accepts": {
      "solution": "sema:Solution#mh:SHA-256:8d55004ffa8f61d6e7d12d588ae4baa9723296dd7fa501759b5e6b1b97e7c9ab"
    },
    "references": {
      "global": "sema:Global#mh:SHA-256:330166f5c16b13500d92d26d0e58ad24657bebccbe295ff8108f363d3539034d",
      "gradient": "sema:Gradient#mh:SHA-256:ecfc2cd2d9eb95778ef04ede8a7e872c33822f174a9e3273fa7bba92b3a30e36",
      "metric": "sema:Metric#mh:SHA-256:aa990383741b05e3b798abc7518d9e65c0cc7832b345d4cfed48714ef2501230",
      "metric_reading": "sema:MetricReading#mh:SHA-256:56a9f18faad70147312f12c1cdb1bcea9cb5d1d61042831d1fe6bfa11414ab26"
    }
  },
  "sema_layer": "Mind",
  "sema_category": "Strategy",
  "used_by": {
    "references": {
      "EvaluatorOptimizer": "sema:EvaluatorOptimizer#mh:SHA-256:ae098f2db7c18e00fa1dea16a305133a0778bd157652d5ee42b65f3bf86eef83",
      "ValuePeg": "sema:ValuePeg#mh:SHA-256:e4409fa4a0947af3497c4c4e059fe147be6a92415c18751aa12baa5983b3fea6"
    },
    "composes_with": {
      "LateralOptimization": "sema:LateralOptimization#mh:SHA-256:168e22fe88101fc8256a1eb40041bda926ae521ccb74f94794aa597cfac32aa4",
      "ManifestPlanning": "sema:ManifestPlanning#mh:SHA-256:8bd4fc5ad5c604cd4814a96d23565a07d890e78800281c81abdfeafe02473896",
      "PUREOptimization": "sema:PUREOptimization#mh:SHA-256:205b2c752f68fe87dde75b65cb848bdad8527dc5713180460e6441eac3b86be6"
    }
  }
}