Policies¶

Decision-making rules that control system behavior.

Overview¶

Policies define how the system makes decisions. They're configurable rules that determine behavior without changing code.

Topic	Description
Task Allocation	Assigning tasks to robots
Station Assignment	Selecting stations for tasks
Batching	Grouping tasks for efficiency
Priority	Task and robot priority rules
Traffic Policies	Congestion and routing decisions

Why Policies?¶

Flexibility¶

Same system, different behaviors:

# Configuration A: Speed-focused
policies:
  task_allocation: nearest_idle

# Configuration B: Balance-focused
policies:
  task_allocation: least_busy

Experimentation¶

Compare approaches easily:

waremax compare scenario.yaml \
  --param policies.task_allocation=nearest_idle \
  --param policies.task_allocation=least_busy

Optimization¶

Find best policy for your warehouse:

waremax sweep scenario.yaml \
  --param "policies.task_allocation=[nearest_idle,least_busy,round_robin]"

Policy Categories¶

Operational Policies¶

Control day-to-day operations:

Policy	Decisions
Task allocation	Which robot gets which task
Station assignment	Which station serves which task
Batching	How to group tasks

Traffic Policies¶

Control robot movement:

Policy	Decisions
Routing	Which path to take
Congestion response	Wait or reroute
Deadlock resolution	How to break deadlocks

Resource Policies¶

Control resource usage:

Policy	Decisions
Charging	When to charge
Maintenance	When to maintain
Station selection	Which charging/maintenance station

Policy Structure¶

Configuration Format¶

policies:
  category:
    policy_name: value
    # or
    policy_name:
      type: complex_type
      param1: value1
      param2: value2

Example¶

policies:
  task_allocation:
    type: nearest_idle
  station_assignment:
    type: shortest_queue
  batching:
    enabled: true
    max_batch_size: 5

Policy Selection¶

Factors to Consider¶

Factor	Consideration
Throughput	Which maximizes tasks/hour
Latency	Which minimizes task time
Fairness	Which balances robot utilization
Robustness	Which handles variability

Trade-offs¶

          Throughput
              ↑
              │     ●A
              │
              │ ●B
              │        ●C
              │
              └──────────────→ Fairness

A: High throughput, low fairness
B: Balanced
C: High fairness, lower throughput

Policy Comparison¶

Using Compare Command¶

waremax compare base.yaml \
  --param policies.task_allocation=nearest_idle \
  --param policies.task_allocation=least_busy

Key Metrics¶

Tasks completed
Average task time
Robot utilization spread
Wait time

Configuration Reference¶

Task Allocation¶

policies:
  task_allocation:
    type: nearest_idle | least_busy | round_robin

Station Assignment¶

policies:
  station_assignment:
    type: nearest | shortest_queue | fastest_completion

Routing¶

policies:
  routing:
    type: shortest_path | congestion_aware
    congestion_weight: 1.5

Batching¶

policies:
  batching:
    enabled: true
    max_batch_size: 5
    max_wait_time_s: 30.0

Best Practices¶

Start Simple¶

Begin with default policies:

policies:
  task_allocation: nearest_idle
  station_assignment: nearest

Measure Impact¶

Always compare before and after:

# Baseline
waremax run baseline.yaml -o baseline_results/

# With new policy
waremax run new_policy.yaml -o new_policy_results/

# Compare
waremax analyze baseline_results/ new_policy_results/

Consider Interactions¶

Policies interact with each other:

Task allocation affects traffic patterns
Traffic patterns affect routing effectiveness
Routing affects task completion times