Table 2 illustrates the SCTTS heuristic algorithm with a step-by-step explanation of the mapping of tasks in a sample workflow in Fig.
1(b) presents the scheduling of the workflow tasks obtained by the SCTTS heuristic algorithm for the sample workflow of Fig.
We compare our proposed heuristic denoted as SCTTS with the MOGA and BE algorithms by conducting a number of experiments, in which we simulate a variety of real-world scenarios.
This section involves a comparison and analysis of the application performance results with objectives such as the makespan, allocation-cost and runtime of the proposed SCTTS algorithm along with the MOGA and the BE algorithms [8, 9].
Obviously, in all instances except when the trade-off factor is zero, the allocation cost of the SCTTS algorithm becomes less than both the BE and MOGA algorithms.
In all cases, except when the trade-off factor is equal to 1, the makespan of the SCTTS algorithm 'is less than both the MOGA and BE algorithms.
6, due to an increase in the application tasks even when it is running 500 tasks, the SCTTS algorithm requires much lower runtime.
7 indicates the allocation cost of the SCTTS algorithm decreases more than 60% compared to the MOGA algorithm due to an increase in the degree of the tasks parallelism in all instances.
9 shows the SCTTS algorithm runtime is at least three orders of magnitude less than the BE and MOGA algorithms in all instances.
9 show, the SCTTS algorithm may deliver a better performance for scheduling the applications with higher degrees of the task parallelism that optimizes the cost-makespan.
10 indicates, with an increase in the number of resources, there is an increase in the allocation cost of the MOGA algorithm compared to the SCTTS algorithm.
11 in corresponding instances, it is understood that in all instances the SCTTS algorithm gives a better performance compared to both MOGA and BE algorithms.