Skip to content

Latest commit

 

History

History
78 lines (59 loc) · 2.68 KB

README.md

File metadata and controls

78 lines (59 loc) · 2.68 KB
Raw

planning-graph

Planning Graph for AI Planning in Python

About

This library implements Planning Graph and its Planner that can be used to solve STRIPS-like AI Planning Problems using PDDL.

When passed the domain and problem representations in PDDL, the library returns Solution Plan if any.

Documentation

If you are interested in understanding the Planning Graph, please read my post at https://towardsdatascience.com/improving-classical-ai-planning-complexity-with-planning-graph-c63d47f87018

Installation

pip install planning-graph

or, you can clone this repository

Example

PDDL for representing planning domain and problem

The Dock-Worker Robots Domain and Problem are provided in the domain directory. You can create your own PDDL files, or you can download from the internet.

Create a Planning Graph from PDDL

from planning_graph.planning_graph import PlanningGraph


planning_graph = PlanningGraph('domain/dock-worker-robot-domain.pddl', 
                               'domain/dock-worker-robot-problem.pddl')

graph = planning_graph.create(max_num_of_levels=10)

planning_graph.create() returns a Graph object if the goal state is achieved, or the maximum number of levels have been reached.

Create visualization

If you want to visualize the generated graph you can set visualize=True, by default it is set to False.

from planning_graph.planning_graph import PlanningGraph


planning_graph = PlanningGraph('domain/dock-worker-robot-domain.pddl', 
                               'domain/dock-worker-robot-problem.pddl',
                               visualize=True)

graph = planning_graph.create(max_num_of_levels=10)
graph.visualize_png("generated_graph.png")

the result looks like the following: alt text

Find a Solution Plan

To find a solution plan you simply have to create a Planner and pass the arguments it requires.

from planning_graph.planning_graph import PlanningGraph
from planning_graph.planning_graph_planner import GraphPlanner


planning_graph = PlanningGraph('domain/dock-worker-robot-domain.pddl', 
                               'domain/dock-worker-robot-problem.pddl')

graph = planning_graph.create(max_num_of_levels=10)
goal = planning_graph.goal
graph_planner = GraphPlanner()
layered_plan = graph_planner.plan(graph, goal)

it returns a layered plan if the solution exists, otherwise, it returns None.

pddlpy

pddlpy included in this repo is the work of Hernán M. Foffani, it is copied from here. It is copied because it won't work when installed as a package due to wrong version of antlr4 package. There are no changes made to it.