Genetic state-grouping algorithm for deep reinforcement learning

Kim, Man-Je; Kim, Jun Suk; Kim, Sungjin James; Kim, Minjung; Ahn, Chang Wook

doi:10.1016/j.eswa.2020.113695

Cited by 9 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The pseudocode of the RLHO method for the TDL model is shown in Algorithm 1, where Q(•,•) refers to the set of state-action values at all time points, and the value of Q t+1 (s t , a t ) is the Q value at time point t + 1. An episode with different number of time steps is one complete play of the agents interacting with the environment in the general reinforcement learning setting [41]. end for 17: end for Figure 1 and Algorithm 1 show the main process of the RLHO method, which can be described as follows:…”

Section: Reinforcement-learning-based Hyperparameter Optimization Met...mentioning

confidence: 99%

A novel transfer deep learning model with reinforcement-learning-based hyperparameter optimization for short-term load forecasting during the COVID-19 pandemic

Zhang

Chen

Zeng

et al. 2022

IFS

View full text Add to dashboard Cite

The coronavirus disease 2019 pandemic has significantly impacted the world. The sudden decline in electricity load demand caused by strict social distancing restrictions has made it difficult for traditional models to forecast the load demand during the pandemic. Therefore, in this study, a novel transfer deep learning model with reinforcement-learning-based hyperparameter optimization is proposed for short-term load forecasting during the pandemic. First, a knowledge base containing mobility data is constructed, which can reflect the changes in visitor volume in different regions and buildings based on mobile services. Therefore, the sudden decline in load can be analyzed according to the socioeconomic behavior changes during the pandemic. Furthermore, a new transfer deep learning model is proposed to address the problem of limited mobility data associated with the pandemic. Moreover, reinforcement learning is employed to optimize the hyperparameters of the proposed model automatically, which avoids the manual adjustment of the hyperparameters, thereby maximizing the forecasting accuracy. To enhance the hyperparameter optimization efficiency of the reinforcement-learning agents, a new advance forecasting method is proposed to forecast the state-action values of the state space that have not been traversed. The experimental results on 12 real-world datasets covering different countries and cities demonstrate that the proposed model achieves high forecasting accuracy during the coronavirus disease 2019 pandemic.

show abstract

Section: Reinforcement-learning-based Hyperparameter Optimization Met...mentioning

confidence: 99%

A novel transfer deep learning model with reinforcement-learning-based hyperparameter optimization for short-term load forecasting during the COVID-19 pandemic

Zhang

Chen

Zeng

et al. 2022

IFS

View full text Add to dashboard Cite

show abstract

“…Several works have ascertained that the GA is more suitable for solving complex and constrained optimization problems in the area of machine learning and data mining [33], [34]. Man-Je et al [35] devised a novel genetic algorithm based on deep reinforcement learning and used it to solve long initial learning times and an overwhelming number of branching factors. Nagar et al [36] adopted a genetic algorithm for efficient feature selection to reduce the dataset dimensions and enhance the classifier pace of a knearest-neighbors technique, which was employed for diagnosing the stage of patients' disease.…”

Section: Genetic Algorithmsmentioning

confidence: 99%

A Novel Many-Objective Recommendation Algorithm for Multistakeholders

Wang

Chen

2020

IEEE Access

View full text Add to dashboard Cite

Most traditional recommender systems focus specifically on increasing consumer satisfaction by providing a list of relevant content to consumers. However, the perspectives of other multisided marketplace stakeholders are also equally important, i.e., the exposure for suppliers or providers and profit for the platform. The suppliers want their products to be presented to users, and the objective of the platform is to maximize their profit. Nevertheless, because consumers' preferences and the objectives of providers as well as the platform may conflict with each other, it degrades the utility of the recommendation methods by only considering users' views. Therefore, in this work, we use a many-objective optimization method to maintain a tradeoff among five objectives for three stakeholders and obtain multiple Pareto front solutions in a single run. We first combine customer lifetime value and user purchase preference to create a new similarity model (Sim_RFMP) to increase the recommendation accuracy of the recommendation list. Furthermore, we propose a many-objective model (NBHXMAOEA) for multistakeholder recommendation. In NBHXMAOEA, we present a novel N-block heuristic crossover operator (NBHX) that recombines blocks of chromosomes based on heuristics. Through extensive experiments, the results demonstrate that our proposed NBHXMAOEA achieves superior performance in terms of average accuracy, diversity, novelty, provider coverage, and platform profit to its competing methods. INDEX TERMS Many-objective, recommender systems, similarity model, stakeholders.

show abstract

“…Various specific implementation strategies of modern heuristic algorithms have been proposed independently, and there are significant differences between them. The modern heuristic algorithms include simulated annealing (SA) algorithm (Duan, 2012;Leite, Melício, and Rosa, 2019;Nino-Ruiz and Yang, 2019), genetic algorithm (GA) (Chekanin and Kulikova, 2017;Kim et al, 2020;Pandey, 2020), Tabu search (TS) algorithm (Ben Abdellafou, Hadda, and Korbaa, 2019;Mohammed and Duffuaa, 2020;Zhang et al, 2020), ant colony algorithm (ACO) (Jalali et al, 2020;Martin et al, 2020;Niu et al, 2007;Srichandum and Pothiya, 2020) and artificial neural network (ANN) (Amari, 1971;Leong et al, 2020). In recent years, an important trend related to computational intelligence has been to establish a new search mechanism based on sports competitions (Bouchekara et al, 2018;Chagwiza et al, 2016;Jaramillo et al, 2016a;Kashan, 2014a;Khattab, Sharieh, and Mahafzah, 2019;Moghdani and Salimifard, 2018;Purnomo and Wee, 2015), through which effective optimization algorithms can be designed.…”

Section: Introductionmentioning

confidence: 99%

A Competitive Volleyball Algorithm to Solve Global Optimization with the Engineering Problems

Sun¹,

Ma²,

Liu³

2021

Preprint

View full text Add to dashboard Cite

In real life, many engineering problems are nonlineble NP problems, in order to solve some of these problems, we put forward a competitive volleyball algorithm.The algorithm proposed in this paper is a meta-heuristic technique based on swarm optimization. It is inspired by the competition between volleyball teams in a league and the improvement in players’ overall abilities in order to win the Most Valuable Player award. Several specific terms relating to competition, such as pre-match reinforcement, single round robin mechanism, optimal strategy constitute the structure of the algorithm. The sensitivity of several parameters of the algorithm is analyzed and tested for three types of benchmark functions: unimodal, high-dimensional multimodal and low-dimensional multimodal functions. Through the use of these three types of test functions, the performance of this algorithm is compared with nine classical metaheuristic algorithms: Genetic Algorithm (GA), Differential Evolution (DE), Harmony Search (HS), Artificial Bee Colony (ABC), Particle Swarm Optimization (PSO), Sine Cosine Algorithm (SCA), Soccer League Competition (SLC), League Championship Algorithm (LCA) and Volleyball Premier League (VPL). CVA has been used to solve three real-world engineering problems. The results show that the performance of the CVA is behaviorally promising and better than the other classical metaheuristic algorithms.

show abstract

Genetic state-grouping algorithm for deep reinforcement learning

Cited by 9 publications

References 11 publications

A novel transfer deep learning model with reinforcement-learning-based hyperparameter optimization for short-term load forecasting during the COVID-19 pandemic

A novel transfer deep learning model with reinforcement-learning-based hyperparameter optimization for short-term load forecasting during the COVID-19 pandemic

A Novel Many-Objective Recommendation Algorithm for Multistakeholders

A Competitive Volleyball Algorithm to Solve Global Optimization with the Engineering Problems

Contact Info

Product

Resources

About