{
  "name": "retail-forecast-optimize-benchmark",
  "version": "1.0.0",
  "records": 10,
  "scenarios": [
    "baseline",
    "promo_spike",
    "holiday_surge",
    "oil_shock",
    "supply_disruption"
  ],
  "schema": [
    "instance_id",
    "scenario_id",
    "oracle_cost",
    "forecast_decoupling",
    "policies.policy",
    "policies.mae",
    "policies.total_cost",
    "policies.service_level",
    "policies.decision_regret"
  ],
  "samples": [
    {
      "instance_id": "baseline_s42",
      "scenario_id": "baseline",
      "seed": 42,
      "oracle_cost": 32649.2,
      "forecast_decoupling": true,
      "insight": "Lower forecast error (quantile_stochastic: MAE 7.01) does not guarantee lower inventory cost. deterministic_p50 wins on total cost ($35,694) despite higher MAE (7.05).",
      "policies": [
        {
          "policy": "naive_ma",
          "mae": 8.48,
          "rmse": 10.837,
          "total_cost": 36618.72,
          "stockout_cost": 17027.16,
          "holding_cost": 515.24,
          "waste_cost": 4530.68,
          "service_level": 0.9124,
          "decision_regret": 3969.5200000000004,
          "runtime_sec": 0.0112
        },
        {
          "policy": "deterministic_p50",
          "mae": 7.051,
          "rmse": 8.402,
          "total_cost": 35694.26,
          "stockout_cost": 44.13,
          "holding_cost": 520.41,
          "waste_cost": 19163.99,
          "service_level": 0.99985,
          "decision_regret": 3045.0600000000013,
          "runtime_sec": 0.0114
        },
        {
          "policy": "quantile_stochastic",
          "mae": 7.012,
          "rmse": 8.351,
          "total_cost": 47380.17,
          "stockout_cost": 0.0,
          "holding_cost": 1053.2,
          "waste_cost": 29993.74,
          "service_level": 1.0,
          "decision_regret": 14730.969999999998,
          "runtime_sec": 0.04
        },
        {
          "policy": "quantile_robust",
          "mae": 7.151,
          "rmse": 8.492,
          "total_cost": 65063.67,
          "stockout_cost": 0.0,
          "holding_cost": 1273.37,
          "waste_cost": 47182.73,
          "service_level": 1.0,
          "decision_regret": 32414.469999999998,
          "runtime_sec": 0.0387
        }
      ]
    },
    {
      "instance_id": "baseline_s123",
      "scenario_id": "baseline",
      "seed": 123,
      "oracle_cost": 32984.4,
      "forecast_decoupling": false,
      "insight": "For this scenario, deterministic_p50 aligns forecast accuracy and decision cost optimality.",
      "policies": [
        {
          "policy": "naive_ma",
          "mae": 9.05,
          "rmse": 11.918,
          "total_cost": 39516.26,
          "stockout_cost": 20061.51,
          "holding_cost": 514.67,
          "waste_cost": 4534.53,
          "service_level": 0.8957999999999999,
          "decision_regret": 6531.860000000001,
          "runtime_sec": 0.0112
        },
        {
          "policy": "deterministic_p50",
          "mae": 6.723,
          "rmse": 8.101,
          "total_cost": 35026.04,
          "stockout_cost": 15.86,
          "holding_cost": 518.22,
          "waste_cost": 18625.78,
          "service_level": 0.9999,
          "decision_regret": 2041.6399999999994,
          "runtime_sec": 0.0115
        },
        {
          "policy": "quantile_stochastic",
          "mae": 6.823,
          "rmse": 8.228,
          "total_cost": 46877.67,
          "stockout_cost": 0.0,
          "holding_cost": 1040.34,
          "waste_cost": 29556.31,
          "service_level": 1.0,
          "decision_regret": 13893.269999999997,
          "runtime_sec": 0.0405
        },
        {
          "policy": "quantile_robust",
          "mae": 6.924,
          "rmse": 8.336,
          "total_cost": 64065.52,
          "stockout_cost": 0.0,
          "holding_cost": 1253.05,
          "waste_cost": 46230.68,
          "service_level": 1.0,
          "decision_regret": 31081.119999999995,
          "runtime_sec": 0.0407
        }
      ]
    },
    {
      "instance_id": "promo_spike_s42",
      "scenario_id": "promo_spike",
      "seed": 42,
      "oracle_cost": 34657.63,
      "forecast_decoupling": true,
      "insight": "Lower forecast error (quantile_stochastic: MAE 10.07) does not guarantee lower inventory cost. deterministic_p50 wins on total cost ($42,309) despite higher MAE (10.11).",
      "policies": [
        {
          "policy": "naive_ma",
          "mae": 10.996,
          "rmse": 14.491,
          "total_cost": 47913.76,
          "stockout_cost": 28341.27,
          "holding_cost": 507.67,
          "waste_cost": 4221.02,
          "service_level": 0.87155,
          "decision_regret": 13256.130000000005,
          "runtime_sec": 0.0097
        },
        {
          "policy": "deterministic_p50",
          "mae": 10.106,
          "rmse": 11.588,
          "total_cost": 42308.81,
          "stockout_cost": 7.76,
          "holding_cost": 593.42,
          "waste_cost": 25357.63,
          "service_level": 0.99995,
          "decision_regret": 7651.18,
          "runtime_sec": 0.0099
        },
        {
          "policy": "quantile_stochastic",
          "mae": 10.068,
          "rmse": 11.574,
          "total_cost": 56397.06,
          "stockout_cost": 0.0,
          "holding_cost": 1241.95,
          "waste_cost": 38655.1,
          "service_level": 1.0,
          "decision_regret": 21739.43,
          "runtime_sec": 0.0404
        },
        {
          "policy": "quantile_robust",
          "mae": 10.264,
          "rmse": 11.763,
          "total_cost": 77531.04,
          "stockout_cost": 0.0,
          "holding_cost": 1509.69,
          "waste_cost": 59271.35,
          "service_level": 1.0,
          "decision_regret": 42873.409999999996,
          "runtime_sec": 0.0403
        }
      ]
    },
    {
      "instance_id": "promo_spike_s123",
      "scenario_id": "promo_spike",
      "seed": 123,
      "oracle_cost": 34334.88,
      "forecast_decoupling": false,
      "insight": "For this scenario, deterministic_p50 aligns forecast accuracy and decision cost optimality.",
      "policies": [
        {
          "policy": "naive_ma",
          "mae": 10.234,
          "rmse": 13.58,
          "total_cost": 45257.79,
          "stockout_cost": 25184.61,
          "holding_cost": 524.65,
          "waste_cost": 4863.42,
          "service_level": 0.88265,
          "decision_regret": 10922.910000000003,
          "runtime_sec": 0.0097
        },
        {
          "policy": "deterministic_p50",
          "mae": 10.224,
          "rmse": 11.816,
          "total_cost": 42383.64,
          "stockout_cost": 0.0,
          "holding_cost": 588.26,
          "waste_cost": 25487.15,
          "service_level": 1.0,
          "decision_regret": 8048.760000000002,
          "runtime_sec": 0.0101
        },
        {
          "policy": "quantile_stochastic",
          "mae": 10.352,
          "rmse": 11.951,
          "total_cost": 56589.58,
          "stockout_cost": 0.0,
          "holding_cost": 1221.91,
          "waste_cost": 38842.67,
          "service_level": 1.0,
          "decision_regret": 22254.700000000004,
          "runtime_sec": 0.0405
        },
        {
          "policy": "quantile_robust",
          "mae": 10.486,
          "rmse": 12.067,
          "total_cost": 77069.18,
          "stockout_cost": 0.0,
          "holding_cost": 1484.65,
          "waste_cost": 58898.49,
          "service_level": 1.0,
          "decision_regret": 42734.299999999996,
          "runtime_sec": 0.0407
        }
      ]
    },
    {
      "instance_id": "holiday_surge_s42",
      "scenario_id": "holiday_surge",
      "seed": 42,
      "oracle_cost": 34844.82,
      "forecast_decoupling": true,
      "insight": "Lower forecast error (quantile_stochastic: MAE 8.71) does not guarantee lower inventory cost. deterministic_p50 wins on total cost ($39,496) despite higher MAE (8.78).",
      "policies": [
        {
          "policy": "naive_ma",
          "mae": 11.593,
          "rmse": 15.689,
          "total_cost": 46639.83,
          "stockout_cost": 28237.1,
          "holding_cost": 514.54,
          "waste_cost": 3043.59,
          "service_level": 0.87475,
          "decision_regret": 11795.010000000002,
          "runtime_sec": 0.0109
        },
        {
          "policy": "deterministic_p50",
          "mae": 8.776,
          "rmse": 10.293,
          "total_cost": 39496.24,
          "stockout_cost": 96.86,
          "holding_cost": 578.36,
          "waste_cost": 22368.4,
          "service_level": 0.99955,
          "decision_regret": 4651.419999999998,
          "runtime_sec": 0.0111
        },
        {
          "policy": "quantile_stochastic",
          "mae": 8.715,
          "rmse": 10.233,
          "total_cost": 53251.34,
          "stockout_cost": 0.0,
          "holding_cost": 1215.24,
          "waste_cost": 35361.11,
          "service_level": 1.0,
          "decision_regret": 18406.519999999997,
          "runtime_sec": 0.0406
        },
        {
          "policy": "quantile_robust",
          "mae": 8.865,
          "rmse": 10.381,
          "total_cost": 74032.72,
          "stockout_cost": 0.0,
          "holding_cost": 1486.12,
          "waste_cost": 55813.79,
          "service_level": 1.0,
          "decision_regret": 39187.9,
          "runtime_sec": 0.0469
        }
      ]
    },
    {
      "instance_id": "holiday_surge_s123",
      "scenario_id": "holiday_surge",
      "seed": 123,
      "oracle_cost": 32296.56,
      "forecast_decoupling": false,
      "insight": "For this scenario, deterministic_p50 aligns forecast accuracy and decision cost optimality.",
      "policies": [
        {
          "policy": "naive_ma",
          "mae": 9.599,
          "rmse": 12.821,
          "total_cost": 40219.91,
          "stockout_cost": 20810.29,
          "holding_cost": 531.37,
          "waste_cost": 4594.88,
          "service_level": 0.89115,
          "decision_regret": 7923.350000000002,
          "runtime_sec": 0.011
        },
        {
          "policy": "deterministic_p50",
          "mae": 7.884,
          "rmse": 9.636,
          "total_cost": 37352.64,
          "stockout_cost": 2.58,
          "holding_cost": 522.3,
          "waste_cost": 20690.4,
          "service_level": 1.0,
          "decision_regret": 5056.079999999998,
          "runtime_sec": 0.0112
        },
        {
          "policy": "quantile_stochastic",
          "mae": 8.001,
          "rmse": 9.74,
          "total_cost": 49488.74,
          "stockout_cost": 0.0,
          "holding_cost": 1063.65,
          "waste_cost": 32069.38,
          "service_level": 1.0,
          "decision_regret": 17192.179999999997,
          "runtime_sec": 0.0412
        },
        {
          "policy": "quantile_robust",
          "mae": 8.113,
          "rmse": 9.873,
          "total_cost": 67092.12,
          "stockout_cost": 0.0,
          "holding_cost": 1281.56,
          "waste_cost": 49176.03,
          "service_level": 1.0,
          "decision_regret": 34795.56,
          "runtime_sec": 0.0409
        }
      ]
    }
  ]
}