The taxicab problem domain is well known in the area of reinforcement learning. Simply put, a taxicab driver is tasked with the problem of picking up a passenger and delivering him to his destination in as few steps as possible. Typically, the taxi is constrained by a limit on the amount of fuel that can be carried.
The canonical taxicab problem is a 5x5 gridworld. There are four cells which serve as possible starting locations and possible destinations for the passenger. There is a...
13 results in 0.0030 seconds.
-
Infinite Mario RL
-
Created by:
Soar
in DomainsInfinite Mario RL
This environment is derived from the Infinite Mario domain from RLCompetition2009 (based on RL-Glue) and uses SML to connect to Soar.
Agents play a variant of Super Mario, a complete side-scrolling video game with destructible blocks, enemies, fireballs, coins, chasms, platforms, etc. The state space is complicated, but factored in an object-oriented way, which captures many aspects of the real world. Challenges include:- Path planning: How can Mario navigate around simple obstacles,
-
Created by:
-
Dice
Dice is a java implementation of a game often called Liar's Dice.
Here is an abbreviated description of the rules used in our implementation of this domain:- Each player gets five six-sided dice and a cup to conceal their dice from other players.
- To begin each round, all players roll their dice under their cups and look at their new 'hand' while keeping it concealed from the other players. The first player begins bidding, picking a quantity of a face number. The quantity states the
-
Water Jug (Reinforcement Learning)
-
Created by:
in AgentsWater Jug (Reinforcement Learning)
This agent demonstrates how to solve the water jug problem using reinforcement learning. It is a modification of the Simple Water Jug Agent. Two main changes are made: templates are used to generate the full space of possible moves that the agent can perform and rewards are set based on the problem solution.
Soar capabilities- Reinforcement learning
- None.
- None.
- The
-
Created by:
-
Taxi (Hierarchical Reinforcement Learning)
-
Created by:
in AgentsTaxi (Hierarchical Reinforcement Learning)
This agent simulates an omniscient taxi driver that uses reinforcement learning and hierarchical task decomposition to improve its performance over runs.
Soar capabilities- Hierarchical task decomposition
- Reinforcement learning
- The agent is included in the Taxi download.
- None.
- Bloch, M.K. Hierarchical Reinforcement Learning in the Taxicab Domain. (Report No. CCA-TR-2009-02). Ann Arbor, MI: Center
-
Created by:
-
Taxi (Reinforcement Learning)
-
Created by:
in AgentsTaxi (Reinforcement Learning)
This agent simulates an omniscient taxi driver that uses reinforcement learning to improve its performance over runs.
Soar capabilities- Reinforcement Learning
- The agent is included in the Taxi download.
- None.
- Bloch, M.K. Hierarchical Reinforcement Learning in the Taxicab Domain. (Report No. CCA-TR-2009-02). Ann Arbor, MI: Center for Cognitive Architecture, University of Michigan. (2009)
-
Created by:
-
Blocks-World (Reinforcement Learning)
-
Created by:
in AgentsBlocks-World (Reinforcement Learning)
This agent contains a version of blocks-world that uses reinforcement learning for move-block, which moves a block to a destination.
Soar capabilities- Reinforcement Learning
- None.
- None.
- The Soar Cognitive Architecture: Chapter 4
- John Laird
- Soar 8,9
- VisualSoar
-
Created by:
-
Blocks-World (Subgoaling with RL)
-
Created by:
in AgentsBlocks-World (Subgoaling with RL)
agent incorporates both operator subgoaling/means ends analysis with reinforcement learning. All search control knowledge (operator evaluation rules) are removed from blocks-world-operator-subgoaling and instead there are RL rules supplemented with rules to compute reward, both in the top state and the substate. Implemented for four blocks.
Soar capabilities- Subgoaling with means-ends analysis
- Reinforcement learning
-
Created by:
-
Blocks-World (Look-Ahead with RL)
-
Created by:
in AgentsBlocks-World (Look-Ahead with RL)
This agent contains a version of blocks-world that involves one level of problem spaces and look-ahead but with two important extensions. It demonstrates how RL-rules can be learned by chunking and then updated in the future. The advantage over simple lookahead is that it doesn't lock on to the one path found during look-ahead after chunking. It will still do some exploration.
Soar capabilities- Reinforcement Learning
- Chunking
- Look-ahead Subgoaling
- Can learn procedural knowledge
-
Created by:
-
Graph Search
-
Created by:
in AgentsGraph Search
An agent that performs graph search using the selection-astar default rules. This approach can be modified to use any of the different selection approaches.
The basic idea behind this agent is that there is a mission to go from place to place (in the mission structure) using go-to-location for each go-to-location use an iterative form of A* search to find the minimal path between each place iteratively select go-to-waypoint to move through the graph
Key data structures (initialized... -
Created by:
-
Infinite Mario RL
-
Created by:
in AgentsInfinite Mario RL
Various agents for learning and playing Infinite Mario domain from RLCompetition 2009. They encode different state representations and learning strategies.
Soar capabilities- Reinforcement learning
- This agent is packaged with the Infinite Mario environment download.
- None.
- None
- Mohan, S. and Laird, J. E. (2011). An Object-Oriented Approach to Reinforcement Learning in an Action Game.
-
Created by:
-
Left-Right Reinforcement Learning Agent
-
Created by:
in AgentsLeft-Right Reinforcement Learning Agent
A very simple agent that demonstrates reinforcement learning. Agent chooses between left and right and must learn that left is preferred.
Soar capabilities- Reinforcement learning
- None.
- None.
- Soar Reinforcement Learning Tutorial
- John Laird
- Soar 9
- VisualSoar
-
Created by:
-
Reinforcement Learning Unit Test
-
Created by:
in AgentsReinforcement Learning Unit Test
An agent that tests four rl-rule sequence cases to behaviorally test the Soar reinforcement learning update mechanism. Learning/discount-rate are set to known values and other parameters can be tweaked. Some of the expected behavior is documented in the README file included.
Soar capabilities- Reinforcement learning
- None.
- None.
- None.
- John Laird
-
Created by: