Search

To address the challenge of partially observable environment states in multi-ship collision avoidance decision-making, a novel collision avoidance decision model is developed based on Partially Observable Markov Decision Processes (POMDP), integrating both static and dynamic obstacles. A new reward mechanism is designed to overcome the problem of sparse rewards, incorporating the international regulations for preventing collisions at sea (COLREGs) into the reward function. The Proximal Policy Optimization (PPO) algorithm is employed to train the model, accompanied by a suitable network structure. Furthermore, image conversion and scaling operations are applied to the network’s architecture to reduce the dimensionality of the state space, thereby enhancing the training speed and fitting performance of the algorithm. Subsequently, a simulation environment is created using the Gym platform, incorporating multiple static and dynamic obstacles, and a series of experiments containing classic encounter scenarios are designed to validate the model. The evaluation metrics, such as the total cumulative reward and training steps, demonstrate that the proposed algorithm is capable of making accurate decisions in an environment with combined static and dynamic obstacles. Through several ablation experiments and analysis of total run-time, it is shown that the proposed reward mechanism effectively addresses the issue of sparse rewards, while the algorithm significantly improves the fitting speed and expedites the training process. The proposed algorithm’s ability to address the problem of partially observable environment states in multi-ship collision avoidance decision-making is confirmed through a comparison with other classical deep reinforcement learning (DRL) algorithms. [ABSTRACT FROM AUTHOR]

CRISPR-Cas systems not only play key roles in prokaryotic acquired immunity, but can also be adapted as powerful genome editing tools. Understanding the native role of CRISPR-Cas systems in providing adaptive immunity can lead to new CRISPR-based technologies. Here, we develop CRISPRminer, a knowledge base and web server to comprehensively collect and investigate the knowledge of CRISPR-Cas systems and generate instructive annotations, including CRISPR arrays and Cas protein annotation, CRISPR-Cas system classification, self-targeting events detection, microbe–phage interaction inference, and anti-CRISPR annotation. CRISPRminer is user-friendly and freely available at http://www.microbiome-bigdata.com/CRISPRminer. Fan Zhang et al. present CRISPRminer, a comprehensive database for exploring CRISPR-Cas systems in more than 3500 microbial species and a web-server for in-depth data mining. CRISPRminer allows researchers to predict CRISPR-Cas systems and investigate self-targeting CRISPRs, microbe-phage interaction networks, and anti-CRISPR features. [ABSTRACT FROM AUTHOR]