In this dissertation, a new efficient routing maintenance algorithm, called Predicting of Future Load-based Routing (PFLR), is introduced for optimizing the routing performance in IP-based networks. The main idea of PFLR algorithm is combing the predicted link load with the current link load with an effective method to optimize the link weights and so reduce the network congestions. Another research objective is introducing a new efficient Traffic Engineering (TE) algorithm, called Prediction-based Decentralized Routing (PDR) algorithm, which is fully decentralized and self-organized approach.
