Mapping interdependent tasks in a computational environment using genetie algorithms

Efficiently processing a large number of tasks is nowadays a necessity, especially when dealing with Big Data or the Internet of Things. Sometimes, there are situations in which tasks cannot be processed until other tasks have finished their execution. This paper takes a look at the possibility of mapping tasks for processing by means of genetic algorithms. Two variations of genetic algorithms are proposed and analyzed; both use a chromosome representation that better suits the considered environment and a fitness function that deals with idle times; the second method employs a targeted mutation operator that speeds up the convergence to a better solution. Simulation tests have shown that better results can be obtained by using genetic algorithms at the expense of a longer algorithm execution time.