Multi-constraint QoS routing using a customized lightweight evolutionary strategy

The ever-increasing transmitting real-time multimedia applications such as VoIP and video conference through the Internet require developing routings methods which guarantee quality of service (QoS) according to the needs of these applications. For these types of applications, a fundamental issue is how to find a feasible path that satisfies multiple constraints. This problem which is known as multi-constraint QoS routing is an NP-complete one, and many research has been devoted to solving it. However, there are still many gaps especially in terms of complexity and speed of the algorithms that must be bridged in order for these methods to be practical. In this regard, in this paper, a novel multi-constraint QoS routing algorithm based on evolutionary strategies is proposed which is lightweight and finds feasible solutions in a very short time. The main reason behind these features is due to the proposed inventive gene decoding mechanism that makes the algorithm needless of any complex evolutionary operators and validation phases. Moreover, a cost function is developed for evaluating the fitness of the potential solutions, which is so flexible for using for as many constraints as required. Therefore, the algorithm is able to search the solution space, no matter how big they are, efficiently and quickly. Simulation results on different network topologies and packet traffics show that our method outperforms competitor algorithms in terms of run time and success ratio, and it is more reliable in different network and traffic scenarios.