Skip to content

Capacity

Resource limits that control robot density.

Concept

Capacity defines how many robots can occupy a resource simultaneously.

Node Capacity

Maximum robots at a location:

traffic:
  node_capacity_default: 1

Edge Capacity

Maximum robots on a path:

traffic:
  edge_capacity_default: 1

Why Capacity Matters

Safety

Prevents physical collisions:

Node capacity = 1
  → Only one robot at each location
  → No collision possible

Realism

Models actual constraints:

  • Aisle widths
  • Intersection sizes
  • Station layouts

Performance

Affects traffic flow:

Higher capacity → More throughput
                → More complex coordination

Node Capacity

Single Occupancy (Default)

Capacity = 1

Before:  [●]  R1 at node

Request: R2 wants node

Result:  R2 waits

Multi-Occupancy

# Intersection with capacity 4
nodes:
  - id: "intersection_1"
    capacity: 4

Capacity = 4

[● ● ●]  R1, R2, R3 at node (3/4)

Request: R4 wants node

Result:  R4 enters (4/4)

Use Cases for Higher Capacity

Location Typical Capacity
Regular aisle 1
Wide intersection 2-4
Waiting area 5-10
Station approach 2-3

Edge Capacity

Single Lane (Default)

Capacity = 1

[N1]═══●═══[N2]
       R1

R2 must wait at N1 or N2

Multi-Lane

edges:
  - from: "N1"
    to: "N2"
    capacity: 2

Capacity = 2

[N1]═══●═══[N2]
       R1
    ═══●═══
       R2

Two robots can traverse simultaneously

Use Cases for Higher Capacity

Path Type Typical Capacity
Regular aisle 1
Wide main aisle 2-3
Highway corridor 3-4

Capacity and Direction

Bidirectional Edges

Same capacity for both directions:

Capacity = 1

[N1]══●══[N2]
      R1 →

R2 ← cannot enter until R1 exits

One-Way Edges

Capacity applies to allowed direction only:

Capacity = 2, one-way →

[N1]══●══[N2]
      R1 →
   ══●══
      R2 →

Both traveling same direction

Capacity Enforcement

Blocking

When at capacity, robots wait:

1. R1 at node N2
2. R2 requests N2
3. N2 at capacity (1/1)
4. R2 waits at current location
5. R1 leaves N2
6. R2 enters N2

Queue Formation

Multiple waiting robots form queues:

[N1]  [N2]  [N3]
 R3    R2    R1
 ←     ←     ●

R1 at destination
R2 waiting for R1
R3 waiting for R2

Capacity Configuration

Global Defaults

traffic:
  node_capacity_default: 1
  edge_capacity_default: 1

Per-Resource Overrides

nodes:
  - id: "main_intersection"
    capacity: 4
  - id: "station_S1_approach"
    capacity: 3

edges:
  - from: "A1"
    to: "A2"
    capacity: 2  # Wide aisle

Capacity Planning

Bottleneck Identification

Low capacity + high traffic = bottleneck:

Traffic: ████████ (high)
Capacity: █ (low)
         ↓
    Bottleneck ⚠️

Solutions

  1. Increase capacity: Widen aisles, expand intersections
  2. Add routes: Parallel paths reduce load
  3. Reduce traffic: Better task distribution

Simulation-Based Tuning

  1. Run simulation with baseline capacity
  2. Identify congestion hotspots
  3. Increase capacity at hotspots
  4. Verify improvement

Effective Throughput

Theoretical Maximum

Max throughput = capacity × 3600 / avg_traversal_time

Example

Edge with: - Capacity: 2 - Length: 6m - Speed: 1.5 m/s - Traversal time: 4s

Max throughput = 2 × 3600 / 4 = 1800 robots/hour

Real Throughput

Always lower due to: - Entry/exit delays - Coordination overhead - Uneven traffic patterns

Best Practices

Conservative Defaults

Start with capacity = 1: - Simplest to reason about - Safest (no collisions) - Increase selectively

Match Physical Layout

3m aisle → capacity = 1
6m aisle → capacity = 2

Test Changes

Small capacity changes can have large effects on congestion patterns.