Differential mutation and novel social learning particle swarm optimization algorithm.

•A Differential mutation and novel Social learning PSO algorithm (DSPSO) is proposed to solve the problems of Social Learning PSO algorithm (SLPSO).•An example-mean learning approach is proposed to obtain a new velocity-updating way.•A single-example learning approach is proposed, and the two approaches are executed alternately to form a novel SLPSO (NSLPSO) to get better exploitation and faster running speed.•A dynamic differential mutation strategy is used to enhance the exploration of NSLPSO.•The experiments are made on the 30- and 50-dimensional complex functions from CEC2013 test set to verify the performance of DSPSO.