In the Internet of Things (IoT) paradigm, all entities in the IoT network, whether home users or industrial things, receive data from other things to make decisions. However, in the decentralized, heterogeneous, and rapidly changing IoT network with billions of devices, deciding about where to get the services or information from is critical, especially because malicious entities can exist in such an unmanaged network. Security provisioning alone cannot solve the issue of service quality or reliability. One way to elevate security and reliability in the IoT network is to bridge the gap of trust between objects, and also between humans and objects, while taking into account the IoT network characteristics. Therefore, a proper trust management system must be established on top of the IoT network service architecture. Trust is related to the manner expected from objects in providing services and recommendations. Recommendations are the basis of decision making in every trust management system. Since trust management ideas in the IoT are still immature, the purpose of this article is to survey, analyze, and compare the approaches that have been taken in building trust management systems for the IoT. We break down the features of such systems by analysis and also do quantitative comparisons by simulation. This article is organized into two main parts. First, studies and approaches in this field are compared from four perspectives: (1) trust computation method, (2) resistance to attacks (3) adherence to the limitations of IoT networks and devices, and (4) performance of the trust management scheme. The second part is quantitative and simulates four major methods in this field and measures their performance. We also make extensive analytical comparisons to demonstrate the similarities and discrepancies of current IoT trust management schemes and extract the essence of a resilient trust management framework.