Skip to content

A ros implementation of distributed optimal transport

License

Notifications You must be signed in to change notification settings

jhughes50/rosdot

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ROSDOT

rosdor_overview

A ROS implementation of Distributed Optimal Transport. There are two types of nodes, target nodes and sources nodes. Sources nodes are supposed to provide target nodes with some resource. This can also be thought of more abstactly, where the algorithm converges to some sort of cost or score between the nodes. ROSDOT will automoatically create topics between source and target nodes in both directions, they take on the name of /source0_to_target0 and /target0_to_source0 and so on for every connection. This repo can be used in two ways.

  1. Run a simulated networked system on a single machine.
  2. Run an individual target or source node.

Requirements

This uses python3, it should work with and ros version that supports python3 and catkin_python_setup. You'll also need scipy and numpy.

Running A Simulation

To run a simulation use the command roslaunch rosdot sim.launch once you've built and sourced the workspace. To add nodes, add an instance of either a target or source node in the launch file and condifure the parameters to your liking. Also, make sure you adjust the network configuration in config/network.yaml such that the number of target and source nodes in correct. You can also adjust the cost parameter between nodes in this file. If you do not want to hard code them you can publish on the /target_cost or /source_cost topics. The cost topics are of a message of type Float64MultiArray.

Running A Single Node

Use the target.launch or the source.launch files to launch a single node, make sure the nodes name is unique in the launch file, and should make the uid parameter. Make sure you configure the network parameters in config/network.yaml as you would for the simulation (see above).

Custom Message Types

Iteration.msg

int64 node_id
float64 data
float64 upper_bound
int64 k

This is the message type used for the iterative updates. node_id is the uid of the target or source node. data is the output of the optimization problem at the iteration, k is the iteration and upper_bound is the upper bound of the node.

TODOs

  • Add in ability to detect convergence rather than using max_iter.
  • Make is stop and wait for new info until computing again
  • Publish the transport scheme on a topic

About

A ros implementation of distributed optimal transport

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published