Modified Krill Herd (MKH) algorithm and its application in dimensional synthesis of a four-bar linkage

Abstract

The paper considers the application of the modified Krill Herd (KH) algorithm for obtaining optimal solutions in dimensional synthesis of a four-bar linkage as a path generator. Certain modifications have been made for the purpose of increasing the performance of the standard KH algorithm for the considered examples of synthesis. In the first modification, besides the initialization of fitness functions, there is also the initialization of the vectors which represent the food location. This modification reflects the actual behavior of krills in their natural environment, i.e. the tendency for optimum swarm density and the best position in relation to the food. The second modification relates to the replacement of the crossover operator with the combination of columns of fitness functions obtained in one iteration. The newly obtained sequence of krills is corrected by the value of physical diffusion, which results in repeated search of the solution space, in the same iteration, in order to improve the optimum found. This is how the Modified Krill Herd (MKH) algorithm tested on four benchmark examples from the synthesis of a four-bar linkage has been obtained. The results obtained by this algorithm confirm its efficiency, i.e. they considerably outperform the results obtained in the cited literature.