Due to high and unpredictable connection delays, privacy gaps, and traffic load of networks connecting cloud computing to end users in many of the Internet of Things (IoT)-based services, some challenges have been created in cloud computing efficiency. Hence, fog computing has been proposed as a solution in order to bring the cloud service closer to the existing things in the ecosystem. Integrating IoT with fog computing is associated with many challenges, including the resource discovery process. In one sense, sensors, devices, and things are the resources in the IoT ecosystem that searching for them regarding the quality of the search and selection can be one of the challenges in the resource discovery process. In the present paper, the hidden Markov chain learning method has been used to cope with this challenge in the IoT ecosystem integrated with the fog computing, to determine the probability of the need for each thing or resource in the near future with the aim of reducing latency and increasing the network use. The simulation in this work has been performed in the Cloudsim platform, and the considered parameters in the proposed method have been compared with TOPSIS, VIKOR and SAW methods.

由于许多基于物联网（IoT）的服务中存在高且不可预测的连接延迟，隐私差距以及将云计算连接到最终用户的网络的流量负载，因此在云计算效率方面带来了一些挑战。因此，已经提出了雾计算作为解决方案，以使云服务更接近生态系统中的现有事物。将物联网与雾计算集成会带来许多挑战，包括资源发现过程。从某种意义上说，传感器，设备和事物是物联网生态系统中的资源，关于搜索和选择质量的搜索可能是资源发现过程中的挑战之一。在本文中，隐藏的马尔可夫链学习方法已被用于应对与雾计算集成的IoT生态系统中的这一挑战，以确定在不久的将来需要每种事物或资源的可能性，目的是减少延迟并增加网络用。在Cloudsim平台上进行了这项工作的仿真，并将所提出的方法中考虑的参数与TOPSIS，VIKOR和SAW方法进行了比较。