Obtaining the accurate value estimation and reducing the estimation bias are the key issues in reinforcement learning. However, current methods that address the overestimation problem tend to introduce underestimation, which face a challenge of precise decision-making in many fields. To address this issue, we conduct a theoretical analysis of the underestimation bias and propose the minmax operation, which allow for flexible control of the estimation bias. Specifically, we select the maximum value of each action from multiple parallel state-action networks to create a new state-action value sequence. Then, a minimum value is selected to obtain more accurate value estimations. Moreover, based on the minmax operation, we propose two novel algorithms by combining Deep Q-Network (DQN) and Double DQN (DDQN), named minmax-DQN and minmax-DDQN. Meanwhile, we conduct theoretical analyses of the estimation bias and variance caused by our proposed minmax operation, which show that this operation significantly improves both underestimation and overestimation biases and leads to the unbiased estimation. Furthermore, the variance is also reduced, which is helpful to improve the network training stability. Finally, we conduct numerous comparative experiments in various environments, which empirically demonstrate the superiority of our method.
- Article type
- Year
Open Access
Full Length Article
Issue
Open Access
Full Length Article
Issue
Recognizing target intent is crucial for making decisions on the battlefield. However, the imperfect and ambiguous character of battlefield situations challenges the validity and causation analysis of classical intent recognition techniques. Facing with the challenge, a target intention causal analysis paradigm is proposed by combining with an Intervention Retrieval (IR) model and a Hybrid Intention Recognition (HIR) model. The target data acquired by the sensors are modelled as Basic Probability Assignments (BPAs) based on evidence theory to create uncertain datasets. Then, the HIR model is utilized to recognize intent for a tested sample from uncertain datasets. Finally, the intervention operator under the evidence structure is utilized to perform attribute intervention on the tested sample. Data retrieval is performed in the sample database based on the IR model to generate the intention distribution of the pseudo-intervention samples to analyze the causal effects of individual sample attributes. The simulation results demonstrate that our framework successfully identifies the target intention under the evidence structure and goes further to analyze the causal impact of sample attributes on the target intention.
京公网安备11010802044758号