Research And Applications On Ant Colony Optimization Based On Multi-Constrained QoS For Wireless Mesh Network

Wireless Mesh Network(WMN) is a multi-hops, self-organizing and self-configuring networkwhich can provide seamless broadband connectivity to network users. With the constinuouslydevelopment of networks and higher users’ demand, it hopes that networks would guarantee the users’Quality of Service(QoS) when providing various applications. However, multi-constrained based QoSrouting is a NP-hard problem which traditional algorithms cannot handle it well. Ant ColonyOptimization(ACO) is a heuristic algorithm inspired by real ants’ behavior that can deal with NP-hardproblems and multi-constrained optimization problems properly. In this paper, ACO is studied andimplemented to apply to multi-constrained QoS routing for WMN. Our work is unfolded mainly fromthe following two aspects:1. It studied the methods of how ACO can be applied to QoS routing problems for wirelessnetworks. On this basis, it investigates the character of the very promising hybrid mesh structure, andproposes an improved ACO routing algorithm, namely IARQM. In IARQM, the degrees of differencebetween the paths found by forward ants are calculated and used to update pheromone in order toeffectively avoid the local pheromone accumulation, which will decrease the search effectiveness.At the same time, considering that the pheromone has significant role in routing, IARQMsupport recovering mechanism to temporarily storage nodes’ accumulated pheromone to improve thealgorithm’s robustness. Moreover, it tends to choose more stable router nodes when routing to furtherensure the stability of the network. Simulated IARQM and compared with other routing algorithms,the results showed that IARQM has better performance and good QoS guarantee ability.2. Generally, single path routing protocol is usually adopted in WMN. However, the single pathrouting protocol has the disadvantage that it may cause network congestion and unbalanced loadproblem. Therefore, on the basis of analyzing the existing ACO based multipath routing algorithms, aparallel perturbation guided ACO multipath routing algorithm based on multi-constrained QoS forWMN(PGARQ) is proposed in this paper.For the sake of generating multipath with low correlation, parallel processing is introduced inPGARQ. It also take advantage of the high calculation and processing speed of the parallel nodes toconstruct paths in parallel. Meanwhile, addressing the drawback of easily trapped into local optima ofACO, an algorithm with combination of ACO and Guided Local Search(GLS) is designed, and a newperturbation method is proposed to optimize the algorithm results. The experiments showed PGARQcould generate multipath with low correlation, improving the calculation accuracy and reduce the calculation time, which in turn ensures the network’s QoS.