Download1 - EJUM

Effective Algorithm for Optimal K-Terminal Reliability of Distributed Systemmethod for three different topologies (ring, bridge, hyper-cube) with eight nodes,respectively.Table 4. The Average Exhaustive Method Execution Time and the ProposedMethod for Three DS Topologies with Eight Nodes.exhaustive metohdthe proposed methodT(s) Lr K request AT(sec) ATlnk(sec) ATT(sec) AT(sec) ATlnk(sec) ATT(sec)4 0.544 0.0083 0.313 0.0062 0.121 0.3260 0.004 0.00594 0.544 0.0083 0.297 0.0052 0.121 0.3205 0.004 0.00584 0.549 0.0083 0.319 0.0062 0.132 0.3333 0.3266 0.005 0.0061 0.0059Ring (n8e8)Bridge (n8e11)(Fig. 3.)0.0~1.00.5~1.00.8~1.00.0~1.04 0.791 0.0113 0.396 0.0072 0.148 0.4451 0.005 0.00794 0.791 0.0113 0.429 0.0082 0.143 0.4542 0.005 0.00804 0.775 0.0113 0.423 0.0082 0.137 0.4451 0.4481 0.005 0.0078 0.00794 1.610 0.0233 0.868 0.0162 0.280 0.9194 0.010 0.01624 1.621 0.0233 0.879 0.0162 0.286 0.9286 0.010 0.01644 1.615 0.0233 0.863 0.0152 0.286 0.9212 0.9231 0.010 0.0162 0.0163Average 0.5659 0.5659 0.5659 0.0100 0.0100 0.0100AT : the seconds of average execution time,AT = ( Σ( execution time))/( totalsimulationcase ) .ATlnk : the seconds of average execution time of same Lr and T(s).ATT : the seconds of average execution time of same T(s).The mean of other notations is described in footnote of Table 2.Hypercube (n8e12)0.5~1.00.8~1.00.0~1.00.5~1.00.8~1.0These data show that the proposed method is more efficient than the exhaustivemethod. In the random method, which obtains the exact solution below 3%, theaverage error from exact solution surpasses 0.26. In our simulation case, the115