My Brain is Full: When More Memory Helps

01/23/2013
by   Christopher Lusena, et al.
0

We consider the problem of finding good finite-horizon policies for POMDPs under the expected reward metric. The policies considered are em free finite-memory policies with limited memory; a policy is a mapping from the space of observation-memory pairs to the space of action-memeory pairs (the policy updates the memory as it goes), and the number of possible memory states is a parameter of the input to the policy-finding algorithms. The algorithms considered here are preliminary implementations of three search heuristics: local search, simulated annealing, and genetic algorithms. We compare their outcomes to each other and to the optimal policies for each instance. We compare run times of each policy and of a dynamic programming algorithm for POMDPs developed by Hansen that iteratively improves a finite-state controller --- the previous state of the art for finite memory policies. The value of the best policy can only improve as the amount of memory increases, up to the amount needed for an optimal finite-memory policy. Our most surprising finding is that more memory helps in another way: given more memory than is needed for an optimal policy, the algorithms are more likely to converge to optimal-valued policies.

READ FULL TEXT
POST COMMENT

Comments

There are no comments yet.

Authors

page 1

page 2

page 3

page 4

page 5

page 6

page 7

page 8

01/23/2013

Solving POMDPs by Searching the Space of Finite Policies

Solving partially observable Markov decision processes (POMDPs) is highl...
09/21/2021

Computing Complexity-aware Plans Using Kolmogorov Complexity

In this paper, we introduce complexity-aware planning for finite-horizon...
09/26/2019

Action Selection for MDPs: Anytime AO* vs. UCT

In the presence of non-admissible heuristics, A* and other best-first al...
06/01/2021

On-Line Policy Iteration for Infinite Horizon Dynamic Programming

In this paper we propose an on-line policy iteration (PI) algorithm for ...
10/07/2019

Stochastic Bandits with Delay-Dependent Payoffs

Motivated by recommendation problems in music streaming platforms, we pr...
04/13/2018

MOVI: A Model-Free Approach to Dynamic Fleet Management

Modern vehicle fleets, e.g., for ridesharing platforms and taxi companie...
06/06/2013

Policy Search: Any Local Optimum Enjoys a Global Performance Guarantee

Local Policy Search is a popular reinforcement learning approach for han...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.