Multi-memetic Mind Evolutionary Computation Algorithm for Loosely Coupled Systems of Desktop Computers

Performance Investigation of Mind Evolutionary Computation Algorithm and Some of Its Modifications

“…Efficiency of a memetic approach to the hybridization of population algorithms was demonstrated in the works [2,4,10]. In this regard, authors also proposed memetic modifications [11,12] of classic MEC, named mMEC and CoMEC, named mCoMEC.…”

“…It was demonstrated in the work [6] that traditional MEC algorithm is bad at optimizing high-dimensional functions, in other words the algorithm isn't suitable for real world problems. In this regard, authors attempted to improve the efficiency of the algorithm without increasing its computational complexity.…”

“…Their popularity is caused by simplicity of implementation and flexibility of application as they are based on the universal idea of evolution. In the meantime, the main disadvantage of such algorithms is their slow convergence to the neighborhood of global optimum because these methods don't utilize any local information about the objective function's landscape [1,2]. This fact often restricts their usage in real world problems, where computation time is a crucial factor.…”

“…The subject of this work is the Mind Evolutionary Computation algorithm (MEC), used by authors as a base algorithm for developing multi-memetic algorithms [2,4]. A concept of the MEC algorithm was firstly proposed in 1998 [5].…”

“…A development of hybrid algorithms implies a combination of various or same methods with different values of free parameters in such a way that advantages of one method would overcome disadvantages of another one. Meta-optimization, on the other hand, implies the adjustment of free parameters' values that would provide the maximum efficiency of an algorithm being investigated [2,3].…”