Path planning and obstacle avoidance are two challenging problems in the study of intelligent robots. In this paper, we develop a new method to alleviate these problems based on deep Q-learning with experience replay ...Path planning and obstacle avoidance are two challenging problems in the study of intelligent robots. In this paper, we develop a new method to alleviate these problems based on deep Q-learning with experience replay and heuristic knowledge. In this method, a neural network has been used to resolve the "curse of dimensionality" issue of the Q-table in reinforcement learning. When a robot is walking in an unknown environment, it collects experience data which is used for training a neural network;such a process is called experience replay.Heuristic knowledge helps the robot avoid blind exploration and provides more effective data for training the neural network. The simulation results show that in comparison with the existing methods, our method can converge to an optimal action strategy with less time and can explore a path in an unknown environment with fewer steps and larger average reward.展开更多
Selecting appropriate resources for running a job efficiently is one of the common objectives in a computational grid. Resource scheduling should consider the specific characteristics of the application, and decide th...Selecting appropriate resources for running a job efficiently is one of the common objectives in a computational grid. Resource scheduling should consider the specific characteristics of the application, and decide the metrics to be used accordingly. This paper presents a distributed resource scheduling framework mainly consisting of a job scheduler and a local scheduler. In order to meet the requirements of different applications, we adopt HGSA, a Heuristic-based Greedy Scheduling Algorithm, to schedule jobs in the grid, where the heuristic knowledge is the metric weights of the computing resources and the metric workload impact factors. The metric weight is used to control the effect of the metric on the application. For different applications, only metric weights and the metric workload impact factors need to be changed, while the scheduling algorithm remains the same. Experimental results are presented to demonstrate the adaptability of the HGSA.展开更多
基金supported by the National Natural Science Foundation of China(61751210,61572441)。
文摘Path planning and obstacle avoidance are two challenging problems in the study of intelligent robots. In this paper, we develop a new method to alleviate these problems based on deep Q-learning with experience replay and heuristic knowledge. In this method, a neural network has been used to resolve the "curse of dimensionality" issue of the Q-table in reinforcement learning. When a robot is walking in an unknown environment, it collects experience data which is used for training a neural network;such a process is called experience replay.Heuristic knowledge helps the robot avoid blind exploration and provides more effective data for training the neural network. The simulation results show that in comparison with the existing methods, our method can converge to an optimal action strategy with less time and can explore a path in an unknown environment with fewer steps and larger average reward.
基金Project supported by the National Natural Science Foundation of China (No. 60225009), and the National Science Fund for Distin-guished Young Scholars, China
文摘Selecting appropriate resources for running a job efficiently is one of the common objectives in a computational grid. Resource scheduling should consider the specific characteristics of the application, and decide the metrics to be used accordingly. This paper presents a distributed resource scheduling framework mainly consisting of a job scheduler and a local scheduler. In order to meet the requirements of different applications, we adopt HGSA, a Heuristic-based Greedy Scheduling Algorithm, to schedule jobs in the grid, where the heuristic knowledge is the metric weights of the computing resources and the metric workload impact factors. The metric weight is used to control the effect of the metric on the application. For different applications, only metric weights and the metric workload impact factors need to be changed, while the scheduling algorithm remains the same. Experimental results are presented to demonstrate the adaptability of the HGSA.
