Questions tagged [planning]

For questions related to the (automated) planning problem, which is the problem of finding a plan, i.e. a sequence of actions to move from an initial state to a goal state or a policy (a function from states to actions), and planning algorithms. There are different ways to define a planning problem (such as PDDL) and solve a planning problem (e.g. GraphPlan). In reinforcement learning, planning consists in finding a policy that solves an MDP.

Filter by
Sorted by
Tagged with
1 vote
0 answers
17 views

Control algorithms when system dynamics are stochastic and/or unknown

I'm working on a traffic signal control problem, which I am currently approaching with Reinforcement Learning, but I want to try some other control algorithms. This is hard for me because we don't ...
Federico Taschin's user avatar
-1 votes
1 answer
333 views

Is my understanding correct regarding the difference between policy and plan?

I am confused regarding the difference between policy and plan in reinforcement learning. According to my understanding, when we calculate the value of state using Bellman equation in deterministic ...
AAA's user avatar
  • 111
0 votes
0 answers
21 views

What is the difference between planning and model-based machine learning?

I am familiar with planning. Given a description of the possible initial states of the world, a description of the desired goals, and a description of a set of possible actions, the planning problem ...
laura.znnt's user avatar
0 votes
1 answer
319 views

How to interpret the output plan of the fast-downward planner

I'm using this domain/problem with the fast-downward planner like this: ./fast-downward.py --plan-file plan.out ../test_domain.pddl ../test_problem.pddl The issue ...
Bilal's user avatar
  • 105
2 votes
0 answers
214 views

What trait of a planning problem makes reinforcement learning a well suited solution?

Planning problems have been the first problems studied at the dawn of AI (Shakey the robot). Graph search (e.g. A*) and planning (e.g. GraphPlan) algorithms can be very efficient at generating a plan. ...
50k4's user avatar
  • 225
1 vote
1 answer
108 views

Is the distribution of state-action pairs from sample based planning accurate for small experience sets?

From the David Silver's lecture 8: Integrating Learning and Planning - based on Sutton and Barto - he talks about using sample-based planning to use our model to take a sample of a state and then use ...
BlueTurtle's user avatar
5 votes
3 answers
512 views

Isn't a simulation a great model for model-based reinforcement learning?

Most reinforcement learning agents are trained in simulated environments. The goal is to maximize performance in (often) the same environment, preferably with a minimum amount of interactions. Having ...
Ray Walker's user avatar
2 votes
1 answer
2k views

Can't solve Towers of Hanoi in PDDL

I'm using PDDL to generate a plan to solve this tower of Hanoi puzzle. I'll give the problem, the rules, the domain and fact sheet for everything. PDDL is telling me that the goal can be simplified ...
Michael's user avatar
  • 131
1 vote
2 answers
410 views

Can two planning PDDL actions be taken simultaneously?

We are discussing planning algorithms currently, and the question is to describe the steps to check if actions could be taken simultaneously. This is a really open-ended question so I'm not sure where ...
dev924's user avatar
  • 11
2 votes
1 answer
349 views

FastDownward PDDL Planner Limitations [closed]

I recently had a look at automated planners and experimented a little bit with FastDownward. As I wanted to start a toy project, I created a PDDL model for the ordinary 3D Rubik's Cube (of course ...
ttttttt1's user avatar
1 vote
0 answers
45 views

Can I solve this assignment problem with RL or AI planning, and if yes how?

I have a list of positive nonzero integers $T=[v_1,\dots,v_𝑛|v_𝑖\in Z^{\neq}]$ which sum up to $V=\sum_i v_i$. Typically, the length of T (number of integers) goes from 100 to 1000. The list is not ...
DeltaIV's user avatar
  • 194
8 votes
1 answer
2k views

MCTS: How to choose the final action from the root

When the time allotted to Monte Carlo tree search runs out, what action should be chosen from the root? The original UCT paper (2006) says bestAction in their ...
user76284's user avatar
  • 347
2 votes
0 answers
97 views

How can I define the relations, preconditions and effects of each operator for the Sokoban puzzle?

I would like to solve the Sokoban puzzle, which consists in moving a character in a 2D map to push boulders into target cells. Each turn, the player can move to an adjacent cell (no diagonals) if it ...
Peyton's user avatar
  • 21
3 votes
1 answer
1k views

Several questions related to UCT and MCTS [closed]

In Bandit Based Monte-Carlo Planning, the article where UCT is introduced as a planning algorithm, there is an algorithm description in page 285 (4 of the pdf). Comparing this implementation of UCT (a ...
Miguel Saraiva's user avatar
2 votes
2 answers
152 views

Can PDDL be utilized for action recognition?

The Planning Domain Definition Language (PDDL) is known for its capabilities of symbolic planning in the state space. A solver will find a sequence of steps to bring the system from a start state to ...
user avatar
3 votes
1 answer
428 views

How to transform a PDDL to search?

I have a question about search and planning: I still haven't understood the difference from the two, but they seem very similar to me; here is a question I am struggling with: "Having formulated a ...
theantomc's user avatar
  • 263
11 votes
3 answers
1k views

What algorithms are considered reinforcement learning algorithms?

What are the areas/algorithms that belong to reinforcement learning? TD(0), Q-Learning and SARSA are all temporal-difference algorithms, which belong to the reinforcement learning area, but is there ...
Miguel Saraiva's user avatar
0 votes
1 answer
1k views

How does a PDDL solver find a solution for a given problem?

As far as I know, in PDDL, an environment is designed as well as the initial state described. When we describe the target state, the solver creates some sort of a graph. How is the graph built and ...
Bryan McGill's user avatar
3 votes
1 answer
2k views

Is there any AI system for finding the best way to schedule university classes?

I was wondering whether there is an AI system which could be used to resolve the class clashes problem which mostly happens in universities. In almost every university students face this problem, ...
Israr Ali's user avatar
  • 103
5 votes
2 answers
5k views

What is "planning" in the context of reinforcement learning, and how is it different from RL and SL?

This is an excerpt taken from Sutton and Barto (pg. 3): Another key feature of reinforcement learning is that it explicitly considers the whole problem of a goal-directed agent interacting with an ...
user avatar
8 votes
2 answers
7k views

What is the difference between search and planning?

I'm reading the book Artificial Intelligence: A Modern Approach (by Stuart Russell and Peter Norvig). However, I don't understand the difference between search and planning. I was more confused when I ...
theantomc's user avatar
  • 263
2 votes
0 answers
199 views

How to choose method for solving planning problems? [closed]

There are many methods and algorithms dealing with planning problems. If I understand correctly, according to Wikipedia, there are classical planning problems, with: a unique known initial state, ...
50k4's user avatar
  • 225
4 votes
1 answer
139 views

What AI technique should I use to assign a person to a task?

I'm trying to learn AI and thinking to apply it to our system. We have an application for the translation industry. What we are doing now is the coordinator $C$ assigns a file to a translator $T$. The ...
Jaime Sangcap's user avatar