A Q-learning algorithm for task scheduling based on improved SVM in wireless sensor networks

Öğrenme Eğrilerinin Karşılaştırılması

“…Son otuz yıl içerisinde öğrenme eğrisinin farklı sektörler tarafından kullanılması ile yeni öğrenme eğrileri önerilmiştir. Önerilen bu öğrenme eğrileri üretim eğrisi modeli [12], karışık öğrenme [13], grup öğrenme [14,15], pozisyon tabanlı öğrenme eğrisi [16,17], insan öğrenme optimizasyon (HLO) algoritması [18][19][20], Q öğrenme [21,22], kuantum modeli [23], belirsiz öğrenme eğrisi [24], kanıta dayalı öğrenme [25], tek seferde öğrenme [26], MICA öğrenme eğrisi [27], bileşen tabanlı öğrenme [28] ve derin öğrenme [28][29][30][31][32] olarak özetlenebilir.…”

Q‐learning based flexible task scheduling in a global view for the Internet of Things

“…For example, the main tasks of sensor devices are event sensing, data sending, data receiving, and so on 24,25 . Because these tasks are some small and simple, they are generally not offloaded to other servers for execution, but they can improve their performance through effective task scheduling 26 . However, there are few researches on effective task scheduling of IoT devices now, which hinders the development of IoT 26 …”

“…There have been a large number of studies on these tasks in WSNs to improve the quality of data collection and extend the lifetime of the network 32‐36 . These studies are actually task scheduling, 26 they regard the operations of sensor nodes as tasks, so the it is meaningful to solve the optimization problem by using the general task scheduling optimization framework 26 . Sensor nodes can not only be organized into a wireless sensor network independently but also be applied to many applications and connected to the Internet in various ways.…”

“…It is an effective method to treat some regular operations of sensor‐based network (such as sleep, event sensing, data sending, and data receiving) as tasks, and use the task scheduling framework to solve the optimization problem. Wei et al 26 proposed the Q‐learning algorithm for task scheduling based on improved support vector machine (ISVM‐Q) scheme for WSNs. The ISVM‐Q scheme embodies Q‐learning algorithm in the following ways: (i) It maps WSN's sensor nodes and environments to the agents and environments of Q‐learning model.…”

“…If performing the operation can bring a positive reward to the node, the node will enhance the probability of performing the operation next time. Otherwise, it will narrow the probability 26 . Therefore, through the ISVM‐Q scheme, each node will perform the operation which maximizes its own profits in WSNs.…”