# Parameter Sweep Configuration
# Comprehensive policy comparison across varied scenarios

[sweep]
simulation_days = 500
policies = ["fifo", "age", "readiness"]

# Dataset Variations
[[datasets]]
name = "baseline"
description = "Default balanced distribution (existing)"
cases = 10000
stage_mix_auto = true  # Use stationary distribution from EDA
urgent_percentage = 0.10
seed = 42

[[datasets]]
name = "admission_heavy"
description = "70% cases in early stages (admission backlog scenario)"
cases = 10000
stage_mix = { "ADMISSION" = 0.70, "ARGUMENTS" = 0.15, "ORDERS / JUDGMENT" = 0.10, "EVIDENCE" = 0.05 }
urgent_percentage = 0.10
seed = 123

[[datasets]]
name = "advanced_heavy"
description = "70% cases in advanced stages (efficient court scenario)"
cases = 10000
stage_mix = { "ADMISSION" = 0.10, "ARGUMENTS" = 0.40, "ORDERS / JUDGMENT" = 0.40, "EVIDENCE" = 0.10 }
urgent_percentage = 0.10
seed = 456

[[datasets]]
name = "high_urgency"
description = "20% urgent cases (medical/custodial heavy)"
cases = 10000
stage_mix_auto = true
urgent_percentage = 0.20
seed = 789

[[datasets]]
name = "large_backlog"
description = "15k cases, balanced distribution (capacity stress test)"
cases = 15000
stage_mix_auto = true
urgent_percentage = 0.10
seed = 999

# Expected Outcomes Matrix (for validation)
# Policy performance should vary by scenario:
# - FIFO: Best fairness, consistent across scenarios
# - Age: Similar to FIFO, slight edge on backlog
# - Readiness: Best efficiency, especially in advanced_heavy and high_urgency