Redirigiendo al acceso original de articulo en 23 segundos...
Inicio  /  Applied Sciences  /  Vol: 12 Par: 23 (2022)  /  Artículo
ARTÍCULO
TITULO

Fresher Experience Plays a More Important Role in Prioritized Experience Replay

Jue Ma    
Dejun Ning    
Chengyi Zhang and Shipeng Liu    

Resumen

Prioritized experience replay (PER) is an important technique in deep reinforcement learning (DRL). It improves the sampling efficiency of data in various DRL algorithms and achieves great performance. PER uses temporal difference error (TD-error) to measure the value of experiences and adjusts the sampling probability of experiences. Although PER can sample valuable experiences according to the TD-error, freshness is also an important character of experiences. It implicitly reflects the potential value of experiences. Fresh experiences are produced by virtue of the current networks and they are more valuable for updating the current network parameters than the past. The sampling of fresh experiences to train the neural networks can increase the learning speed of the agent, but few algorithms can perform this job efficiently. To solve this issue, a novel experience replay method is proposed in this paper. We first define that the experience freshness is negatively correlated with the number of replays. A new hyper-parameter, the freshness discounted factor ?? µ , is introduced in PER to measure the experience freshness. Further, a novel experience replacement strategy in the replay buffer is proposed to increase the experience replacement efficiency. In our method, the sampling probability of fresh experiences is increased by raising its priority properly. So the algorithm is more likely to choose fresh experiences to train the neural networks during the learning process. We evaluated this method in both discrete control tasks and continuous control tasks via OpenAI Gym. The experimental results show that our method achieves better performance in both modes of operation.

 Artículos similares

       
 
William Villegas-Ch and Joselin García-Ortiz    
In the digital age, the personalization of learning has become a critical priority in education. This article delves into the cutting-edge of educational innovation by exploring the essential role of ontology-based knowledge representation in transformin... ver más
Revista: Computers

 
Sara Sanz Sáenz, Gabriel Diaz-Hernandez, Lutz Schweter and Pieter Nordbeck    
The size of container vessels is continuously growing, always exceeding expectations. Port authorities and terminals need to constantly adapt and face challenges related to maritime infrastructure, equipment, and operations, as these are the principal ar... ver más

 
Mlamuleli Tshuma, Johannes A. Belle and Alice Ncube    
Communities around the world living in either urban or rural areas continue to experience serious WASH problems during flood episodes. Communities and individual households are affected differently depending on their coping capacities and their resource ... ver más
Revista: Water

 
Xuepeng Liu, Shiping He and Jianhua Zheng    
Deep-sea pressure vessels are specialized pressure vessels designed for automatic deployment from underwater to the surface. These vessels find extensive applications in underwater life-saving and transportation. Their interiors are furnished with a pair... ver más
Revista: Applied Sciences

 
Ming-Kuo Chen, Hsin-Nan Chien and Ruo-Lan Liu    
This study explored college students? learning experiences and outcomes as club committee members. Using a linear regression model, it investigated the relevance of personal background variables and club learning experiences to club learning outcomes. Th... ver más