Evolutionary Computation: Techniques and Applications by Ashish M. Gujarathi & B. V. Babu

Author:Ashish M. Gujarathi & B. V. Babu
Language: eng
Format: mobi
Publisher: Apple Academic Press
Published: 2016-11-30T23:00:00+00:00

where is the component in dimension d of the i-th particle velocity in iteration t, is the component in dimension d of the i-th particle position in iteration t, c1, c2 are constant weight factors, pi is the best position achieved by particle i, pgd is the best position found by the neighbors of particle i, r1, r2 are random factors in the [0,1] interval, and w is the inertia weight.

10.4 SIMULATION RESULTS

We evaluate the performance of the path based and tree-based evolutionary algorithms for QoS multicast routing by implementing these algorithms in Visual C++. The experiments are performed on an Intel Core i3 @ 2.27 G.Hz. and 4 GB RAM based platform running Windows 7.0. The nodes are positioned randomly in an area of size 4000 km x 2400 km. The Euclidean metric is then used to determine the distance between each pair of nodes. The network topology used in our simulation was generated randomly using Waxman’s topology [35]. The edges are introduced between the pairs of nodes u, v with a probability that depends on the distance between them. The edge probability is given by p(u, v) = β exp(– l(u, v)/αL), where l(u,v) is the Euler distance from node u to v and L is the maximum distance between any two points in the network. The delay, loss rate, bandwidth and cost of the links are set randomly from 1 to 30, 0.0001 to 0.01, 2 to 10 Mbps and 1 to 100, respectively.

The source node is selected randomly and destination nodes are picked up uniformly from the set of nodes chosen in the network topology. The delay bound, the delay jitter bound and the loss bound are set 120 ms, 60 ms and 0.05, respectively. The bandwidth requested by a multicast application is generated randomly. We generate 30 multicast trees randomly to study and compare the performance of the tree-based algorithms. The simulation is run for 100 times for each case and the average of the multicast tree cost is taken as the output. We also generate 20 shortest paths for each destination to study and compare the performance of the path-based algorithms. We vary the network size from 20 to 140 and the number of destinations is considered as 20% of the number of nodes in the network. The performance of these algorithms is studied in terms of multicast tree cost, delay and delay-jitter.

The Figure 10.6 shows the comparison of multicast tree cost of various path based algorithms such as BFPSO [18], QPSO [24], GAPSO [1], PSO [26] with respect to varying network size .The results show that the BF-PSO performs better than QPSO, PSO and GA-PSO in terms of cost. This is because the BFPSO uses an efficient loop deletion procedure to generate a better-multicast tree while combining the paths to the destinations. Furthermore, the BFO has a powerful searching capability to select the optimal set of paths to generate the best possible tree. The convergence speed of BFO has also been improved with PSO.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.