Route13

Route13 is a framework for building simulators and optimizers for transportation networks. Route13 includes a number of naive, brute-force and heuristics based optimizers, but its pluggable architecture allows the use of more sophisticated optimizers, such as linear programming solvers and ML models. Route13 scenarios include forklifts in warehouses, baggage carts at airports, and trucks on highways. Basically anything that involves workers or equipment moving loads over a network while satisfying constraints around delivery times, equipment capacities, and worker schedules.

For information on how Route13 works, please see our design documents.

Try Route13

Route13 is a Node.js project, written in TypeScript. In order to use Route13 you must have Node installed on your machine. Route13 has been tested with Node version 9.4.0.

Route13 will be available as an npm package. To install Route13,

% npm install route13

To run the samples, it is best to build Route13 from sources. First, clone the repo . Then run the following commands:

% npm install
% npm run compile

Unit tests are based on Mocha and Chai and can be run with

% npm run test

Simulator Sample

This sample models a simple environment with 11 locations, uniformly spaced along a line. At time zero, there are 3 carts, each positioned at location 0. We start the simulation with the following jobs:

• Job 0: Move 5 items from location 2 to 10 between the times 300 and 3000.
• Job 1: Move 5 items from location 3 to 4 between the times 300 and 3000.
• Job 2: Go out of service at location 7 between the times 3000 and 4000.
• Job 3: Move 7 items from location 8 to 4 between the times 1300 and 3000.

This simulator makes use of a Dispatcher that is configured to assign jobs randomly to carts as the carts become available.

The simulator is configured with a TextTrace object that logs status to the console.

% node build/samples/simulator-demo.js

0: Job 0 assigned to cart 0.
0: Cart 0 departs location 0 for location 2.
0: Job 3 assigned to cart 1.
0: Cart 1 departs location 0 for location 8.
0: Job 1 assigned to cart 2.
0: Cart 2 departs location 0 for location 3.
100: Cart 2 passes location 1.
100: Cart 1 passes location 1.
100: Cart 0 passes location 1.
200: Cart 1 passes location 2.
200: Cart 0 arrives at location 2.
200: Cart 0 waits until 300.
200: Cart 2 passes location 2.
300: Cart 0 begins loading 5 items.
300: Cart 2 arrives at location 3.
300: Cart 2 begins loading 5 items.
300: Cart 1 passes location 3.
325: Cart 2 finishes loading (payload=5).
325: Cart 2 departs location 3 for location 4.
325: Cart 0 finishes loading (payload=5).
325: Cart 0 departs location 2 for location 10.
400: Cart 1 passes location 4.
425: Cart 0 passes location 3.
425: Cart 2 arrives at location 4.
425: Cart 2 begins unloading 5 items.
435: Cart 2 finishes unloading (payload=0).
435: Job 1 succeeded.
435: Job 2 assigned to cart 2.
435: Cart 2 departs location 4 for location 7.
500: Cart 1 passes location 5.
525: Cart 0 passes location 4.
535: Cart 2 passes location 5.
600: Cart 1 passes location 6.
625: Cart 0 passes location 5.
635: Cart 2 passes location 6.
700: Cart 1 passes location 7.
725: Cart 0 passes location 6.
735: Cart 2 arrives at location 7.
735: Cart 2 suspends service at location 7.
735: Cart 2 waits until 4000.
800: Cart 1 arrives at location 8.
800: Cart 1 waits until 1300.
825: Cart 0 passes location 7.
925: Cart 0 passes location 8.
1025: Cart 0 passes location 9.
1125: Cart 0 arrives at location 10.
1125: Cart 0 begins unloading 5 items.
1135: Cart 0 finishes unloading (payload=0).
1135: Job 0 succeeded.
1300: Cart 1 begins loading 7 items.
1335: Cart 1 finishes loading (payload=7).
1335: Cart 1 departs location 8 for location 4.
1435: Cart 1 passes location 7.
1535: Cart 1 passes location 6.
1635: Cart 1 passes location 5.
1735: Cart 1 arrives at location 4.
1735: Cart 1 begins unloading 7 items.
1749: Cart 1 finishes unloading (payload=0).
1749: Job 3 succeeded.
4000: Job 2 succeeded.
4000: Cart 2 resumes service at location 7.
Simulation ended.

Route Planning Sample

This sample demonstrates route planning for a single Cart. In this case the Dispatcher has assigned the following jobs to a Cart:

• Job 0: Move 5 items from location 2 to 10 between the times 300 and 3000.
• Job 1: Move 5 items from location 3 to 4 between the times 300 and 3000.
• Job 2: Go out of service at location 7 between the times 3000 and 4000.

We would like to determine ordering of pickups and dropoffs that minimizes working time.

The RoutePlanner has been configured to print out those plans it considered, but rejected because they violated constraints. The sample prints the optimal plan at the end.

=================
Failed:

Plan for cart 0 (working time = 2260):
PICKUP 5 bags at gate 2 after 300 (job 0)
    0: drive for 200s to gate 2
    200: wait 100s until 300
    300: load 5 bags in 25s.
    325: completed
DROPOFF 5 bags at gate 10 before 3000 (job 0)
    325: drive for 800s to gate 10
    1125: unload 5 bags in 10s.
    1135: completed
PICKUP 5 bags at gate 3 after 300 (job 1)
    1135: drive for 700s to gate 3
    1835: load 5 bags in 25s.
    1860: completed
SUSPEND at gate 7 before 3000 until 4000 (job 2)
    1860: drive for 400s to gate 7
    2260: suspend operations
    2260: wait 1740s until 4000
    4000: resume operations
    4000: completed
DROPOFF 5 bags at gate 4 before 3000 (job 1)
    4000: drive for 300s to gate 4
    4300: unload 5 bags in 10s.
    4310: CONTRAINT VIOLATED - dropoff after deadline 3000

=================
Failed:

Plan for cart 0 (working time = 1860):
PICKUP 5 bags at gate 2 after 300 (job 0)
    0: drive for 200s to gate 2
    200: wait 100s until 300
    300: load 5 bags in 25s.
    325: completed
DROPOFF 5 bags at gate 10 before 3000 (job 0)
    325: drive for 800s to gate 10
    1125: unload 5 bags in 10s.
    1135: completed
SUSPEND at gate 7 before 3000 until 4000 (job 2)
    1135: drive for 300s to gate 7
    1435: suspend operations
    1435: wait 2565s until 4000
    4000: resume operations
    4000: completed
PICKUP 5 bags at gate 3 after 300 (job 1)
    4000: drive for 400s to gate 3
    4400: load 5 bags in 25s.
    4425: completed
DROPOFF 5 bags at gate 4 before 3000 (job 1)
    4425: drive for 100s to gate 4
    4525: unload 5 bags in 10s.
    4535: CONTRAINT VIOLATED - dropoff after deadline 3000

=================
Failed:

... 13 more failed plans ...

=================
Failed:

Plan for cart 0 (working time = 1125):
SUSPEND at gate 7 before 3000 until 4000 (job 2)
    0: drive for 700s to gate 7
    700: suspend operations
    700: wait 3300s until 4000
    4000: resume operations
    4000: completed
PICKUP 5 bags at gate 3 after 300 (job 1)
    4000: drive for 400s to gate 3
    4400: load 5 bags in 25s.
    4425: completed
DROPOFF 5 bags at gate 4 before 3000 (job 1)
    4425: drive for 100s to gate 4
    4525: unload 5 bags in 10s.
    4535: CONTRAINT VIOLATED - dropoff after deadline 3000

Considered 16 failed plans.
Considered 3 successful plans.

#########################
Planning Complete

Plan for cart 0 (working time = 1470):
PICKUP 5 bags at gate 2 after 300 (job 0)
    0: drive for 200s to gate 2
    200: wait 100s until 300
    300: load 5 bags in 25s.
    325: completed
PICKUP 5 bags at gate 3 after 300 (job 1)
    325: drive for 100s to gate 3
    425: load 5 bags in 25s.
    450: completed
DROPOFF 5 bags at gate 4 before 3000 (job 1)
    450: drive for 100s to gate 4
    550: unload 5 bags in 10s.
    560: completed
DROPOFF 5 bags at gate 10 before 3000 (job 0)
    560: drive for 600s to gate 10
    1160: unload 5 bags in 10s.
    1170: completed
SUSPEND at gate 7 before 3000 until 4000 (job 2)
    1170: drive for 300s to gate 7
    1470: suspend operations
    1470: wait 2530s until 4000
    4000: resume operations
    4000: completed

Graph Sample

The Graph class uses the Floyd-Warshall algorithm to compute travel times and paths through a directed graph with weighted edges. The Graph.cost() and Graph.next() methods can be used to supply the TransitTimeEstimator and RouteNextStep parameters to the Environment constructor.

This sample demonstrates the algorithm on a small graph with the following edges:

• { from: 0, to: 2, weight: -2}
• { from: 1, to: 0, weight: 4}
• { from: 1, to: 2, weight: 3}
• { from: 2, to: 3, weight: 2}
• { from: 3, to: 1, weight: -1}

The sample outputs the Graph.distances matrix, the Graph.next matrix, and the shortest paths between all pairs of vertices.

% node build/samples/graph-demo.js

=== Iteration 0 ===
Distances
0: 0, -1, -2, 0
1: 4, 0, 2, 4
2: 5, 1, 0, 2
3: 3, -1, 1, 0

Next
0: x 2 2 2
1: 0 x 0 0
2: 3 3 x 3
3: 1 1 1 x

0 => 0: []
0 => 1: [2,3,1]
0 => 2: [2]
0 => 3: [2,3]
1 => 0: [0]
1 => 1: []
1 => 2: [0,2]
1 => 3: [0,2,3]
2 => 0: [3,1,0]
2 => 1: [3,1]
2 => 2: []
2 => 3: [3]
3 => 0: [1,0]
3 => 1: [1]
3 => 2: [1,0,2]
3 => 3: []