Abstract:  A QoS link selection algorithm based on multi-constraint condition feedback is proposed for the link
selection in the ubiquitous positioning system of Internet of Things to satisfy the requirement of multiple
restrictions. By using the characteristic of wireless communication broadcasting, the problem of the multi-
constraint service quality routing is transformed into the problem of minimum cost in satisfying the link length,
bandwidth, delay, delay jitter, packet loss rate, and cost and so on. In this algorithm, the multi-hop
autonomous link model of Internet of Things is established and the target energy function is defined, and the
Hopfield neural network model is used to transform multiple constraint conditions into a polynomial extremal
problem. The algorithm avoids the problem of selecting a large number of parameters without changing the system
link selection strategy. It realizes the automatic selection of link and satisfies the requirement of real-time
communication.

Key words:  the quality of service (QoS) link selection algorithm, hopfield artificial neural networks (ANNs), internet of things, network service, the shortest path