Skip to content

Policy Comparisons

Examples comparing different policy configurations.

Task Allocation Comparison

Compare how different task allocation policies affect performance.

Setup

# base_scenario.yaml (common configuration)
simulation:
  duration_s: 3600

robots:
  count: 10
  speed_m_s: 1.5

stations:
  - { id: S1, node: 30, type: pick, concurrency: 2 }
  - { id: S2, node: 31, type: pick, concurrency: 2 }

orders:
  generation:
    type: poisson
    rate_per_hour: 300

Run Comparison

waremax compare base_scenario.yaml \
  --param policies.task_allocation=nearest_idle \
  --param policies.task_allocation=least_busy \
  --param policies.task_allocation=round_robin \
  --runs 5

Expected Results

Policy Throughput Utilization Spread Avg Travel
nearest_idle 295/hr 15% 12.3s
least_busy 280/hr 5% 15.8s
round_robin 265/hr 2% 18.2s

Analysis

  • nearest_idle: Best throughput, some robots work more than others
  • least_busy: Balanced workload, slightly longer travel
  • round_robin: Most balanced but ignores efficiency

Station Assignment Comparison

Compare station selection strategies.

Setup

# Unbalanced stations
stations:
  - { id: S1, node: 30, type: pick, concurrency: 1 }  # Closer
  - { id: S2, node: 31, type: pick, concurrency: 2 }  # Further but larger

Run Comparison

waremax compare base_scenario.yaml \
  --param policies.station_assignment=nearest \
  --param policies.station_assignment=shortest_queue \
  --param policies.station_assignment=fastest_completion

Expected Results

Policy Throughput S1 Util S2 Util Avg Queue
nearest 250/hr 95% 45% 5.2
shortest_queue 285/hr 72% 78% 2.1
fastest_completion 295/hr 68% 82% 1.8

Analysis

  • nearest: Creates imbalance, S1 overloaded
  • shortest_queue: Better balance
  • fastest_completion: Best throughput, considers all factors

Routing Policy Comparison

Compare pathfinding strategies under congestion.

Setup

robots:
  count: 20  # Higher density for congestion

traffic:
  congestion_response: reroute

Run Comparison

waremax compare congestion_scenario.yaml \
  --param routing.policy=shortest_path \
  --param "routing.policy=congestion_aware,routing.congestion_weight=1.0" \
  --param "routing.policy=congestion_aware,routing.congestion_weight=2.0"

Expected Results

Policy Throughput Avg Wait Avg Travel Deadlocks
shortest_path 380/hr 12.5s 14.2s 3
congestion_aware (1.0) 420/hr 6.8s 16.1s 0
congestion_aware (2.0) 410/hr 4.2s 18.5s 0

Analysis

  • shortest_path: Direct routes but gets congested
  • congestion_aware (1.0): Good balance
  • congestion_aware (2.0): Very few waits but longer paths

Batching Comparison

Compare batching vs. individual task execution.

Setup

orders:
  generation:
    type: poisson
    rate_per_hour: 400

Run Comparison

waremax compare base_scenario.yaml \
  --param policies.batching.enabled=false \
  --param "policies.batching.enabled=true,policies.batching.max_batch_size=3" \
  --param "policies.batching.enabled=true,policies.batching.max_batch_size=5"

Expected Results

Config Throughput Travel/Task Task Latency
No batching 380/hr 15.2s 42s
Batch 3 410/hr 11.8s 48s
Batch 5 430/hr 9.5s 55s

Analysis

  • Batching improves throughput but increases individual task latency
  • Batch size 3-5 is typically optimal
  • Trade-off depends on latency requirements

Combined Policy Optimization

Find the best combination of policies.

Sweep Configuration

waremax sweep base_scenario.yaml \
  --param "policies.task_allocation=[nearest_idle,least_busy]" \
  --param "policies.station_assignment=[shortest_queue,fastest_completion]" \
  --param "routing.congestion_weight=[1.0,1.5,2.0]" \
  --runs 3

Top Configurations

Rank Allocation Station Weight Throughput
1 nearest_idle fastest_completion 1.5 435/hr
2 nearest_idle fastest_completion 1.0 428/hr
3 nearest_idle shortest_queue 1.5 420/hr
4 least_busy fastest_completion 1.5 415/hr 
policies:
  task_allocation: nearest_idle
  station_assignment: fastest_completion

routing:
  policy: congestion_aware
  congestion_weight: 1.5

Running Your Own Comparisons

Template Script

#!/bin/bash
# policy_comparison.sh

BASE_SCENARIO="scenario.yaml"
OUTPUT_DIR="comparison_results"

# Run comparisons
waremax compare $BASE_SCENARIO \
  --param policies.task_allocation=nearest_idle \
  --param policies.task_allocation=least_busy \
  --runs 5 \
  -o $OUTPUT_DIR/task_allocation/

# Analyze
waremax analyze $OUTPUT_DIR/task_allocation/

# Generate report
waremax analyze $OUTPUT_DIR/task_allocation/ \
  --export comparison_report.csv