M-Learning: Heuristic Approach for Delayed Rewards in Reinforcement Learning

The current design of reinforcement learning methods requires extensive computational resources. Algorithms such as Deep Q-Network (DQN) have obtained outstanding results in advancing the field. However, the need to tune thousands of parameters and run millions of training episodes remains a signifi...

Full description

Saved in:
Bibliographic Details
Main Authors: Cesar Andrey Perdomo Charry, Marlon Sneider Mora Cortes, Oscar J. Perdomo
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Mathematics
Subjects:
Online Access:https://www.mdpi.com/2227-7390/13/13/2108
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849319959653515264
author Cesar Andrey Perdomo Charry
Marlon Sneider Mora Cortes
Oscar J. Perdomo
author_facet Cesar Andrey Perdomo Charry
Marlon Sneider Mora Cortes
Oscar J. Perdomo
author_sort Cesar Andrey Perdomo Charry
collection DOAJ
description The current design of reinforcement learning methods requires extensive computational resources. Algorithms such as Deep Q-Network (DQN) have obtained outstanding results in advancing the field. However, the need to tune thousands of parameters and run millions of training episodes remains a significant challenge. This document proposes a comparative analysis between the Q-Learning algorithm, which laid the foundations for Deep Q-Learning, and our proposed method, termed M-Learning. The comparison is conducted using Markov Decision Processes with the delayed reward as a general test bench framework. Firstly, this document provides a full description of the main challenges related to implementing Q-Learning, particularly concerning its multiple parameters. Then, the foundations of our proposed heuristic are presented, including its formulation, and the algorithm is described in detail. The methodology used to compare both algorithms involved training them in the Frozen Lake environment. The experimental results, along with an analysis of the best solutions, demonstrate that our proposal requires fewer episodes and exhibits reduced variability in the outcomes. Specifically, M-Learning trains agents 30.7% faster in the deterministic environment and 61.66% faster in the stochastic environment. Additionally, it achieves greater consistency, reducing the standard deviation of scores by 58.37% and 49.75% in the deterministic and stochastic settings, respectively. The code will be made available in a GitHub repository upon this paper’s publication.
format Article
id doaj-art-c59857e92db645a0b9a8a0ffcc1fc06c
institution Kabale University
issn 2227-7390
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Mathematics
spelling doaj-art-c59857e92db645a0b9a8a0ffcc1fc06c2025-08-20T03:50:16ZengMDPI AGMathematics2227-73902025-06-011313210810.3390/math13132108M-Learning: Heuristic Approach for Delayed Rewards in Reinforcement LearningCesar Andrey Perdomo Charry0Marlon Sneider Mora Cortes1Oscar J. Perdomo2Faculty of Engineering, Universidad Distrital Francisco José de Caldas, Bogotá 111611, ColombiaFaculty of Engineering, Universidad Distrital Francisco José de Caldas, Bogotá 111611, ColombiaDeptartment of Electrical and Electronic, Universidad Nacional de Colombia, Bogotá 111321, ColombiaThe current design of reinforcement learning methods requires extensive computational resources. Algorithms such as Deep Q-Network (DQN) have obtained outstanding results in advancing the field. However, the need to tune thousands of parameters and run millions of training episodes remains a significant challenge. This document proposes a comparative analysis between the Q-Learning algorithm, which laid the foundations for Deep Q-Learning, and our proposed method, termed M-Learning. The comparison is conducted using Markov Decision Processes with the delayed reward as a general test bench framework. Firstly, this document provides a full description of the main challenges related to implementing Q-Learning, particularly concerning its multiple parameters. Then, the foundations of our proposed heuristic are presented, including its formulation, and the algorithm is described in detail. The methodology used to compare both algorithms involved training them in the Frozen Lake environment. The experimental results, along with an analysis of the best solutions, demonstrate that our proposal requires fewer episodes and exhibits reduced variability in the outcomes. Specifically, M-Learning trains agents 30.7% faster in the deterministic environment and 61.66% faster in the stochastic environment. Additionally, it achieves greater consistency, reducing the standard deviation of scores by 58.37% and 49.75% in the deterministic and stochastic settings, respectively. The code will be made available in a GitHub repository upon this paper’s publication.https://www.mdpi.com/2227-7390/13/13/2108reinforcement learningexploration–exploitation dilemmaQ-learningfrozen lakeheuristic approach
spellingShingle Cesar Andrey Perdomo Charry
Marlon Sneider Mora Cortes
Oscar J. Perdomo
M-Learning: Heuristic Approach for Delayed Rewards in Reinforcement Learning
Mathematics
reinforcement learning
exploration–exploitation dilemma
Q-learning
frozen lake
heuristic approach
title M-Learning: Heuristic Approach for Delayed Rewards in Reinforcement Learning
title_full M-Learning: Heuristic Approach for Delayed Rewards in Reinforcement Learning
title_fullStr M-Learning: Heuristic Approach for Delayed Rewards in Reinforcement Learning
title_full_unstemmed M-Learning: Heuristic Approach for Delayed Rewards in Reinforcement Learning
title_short M-Learning: Heuristic Approach for Delayed Rewards in Reinforcement Learning
title_sort m learning heuristic approach for delayed rewards in reinforcement learning
topic reinforcement learning
exploration–exploitation dilemma
Q-learning
frozen lake
heuristic approach
url https://www.mdpi.com/2227-7390/13/13/2108
work_keys_str_mv AT cesarandreyperdomocharry mlearningheuristicapproachfordelayedrewardsinreinforcementlearning
AT marlonsneidermoracortes mlearningheuristicapproachfordelayedrewardsinreinforcementlearning
AT oscarjperdomo mlearningheuristicapproachfordelayedrewardsinreinforcementlearning