A number of experiments are designed in this section to do the performance and delay comparisons for the MRDCM trees, which are generated by applying the load-based DFS channel assignment procedure to the multicast trees constructed by the LMCM, the load-based greedy and the shortest-path (SP) algorithms.
10, the performance ratio [THETA] of LMCM method is consistently superior to the other two methods for various destination ratios in different sizes of networks when the path delay is set to be 15.
11 where the simulations for LMCM algoithm on 100-node networks is tested.
12, the average value of performance ratios generated by the CA with all the channels available is about twice as large as that generated by the CA with orthogonal channels only for the LMCM algorithm.
13, the same set of multicast trees built by LMCM is used for allocating conflict-free channels based on two different CA procedures proposed in section 4.3.3.
14, the average delay of MRDCM trees produced by the SP algorithm is the smallest, whereas the average delay of the LMCM method is the largest and very close to that of greedy method for various destination ratios for 100-node networks.
This phenomenon is due to that the number of serviced users found by LMCM algorithm is much larger than that found by SP method.
According to the computation complexity analysis shown in the section 4, the running time of the LMCM and load-based greedy algorithms is O([[absolute value of V].sup.2] x log [absolute value of V] xL) and O([[absolute value of V].sup.2] L) respectively.
For each set of simulations, the summation of 100 runs CPU time is used for comparing executing efficiency for SP, load-based greedy and LMCM algorithms.