Figure 4 Probabilistic search toward non-dominated (elite) solutions in each iteration with a dynamic decreasing learning (neighborhood) ratio. Fig. 3. The shaking trajectory-based search concepts: (a) A neuron’s center search along the elite solutions. Dominated chromosomes are replaced with new ones in « trajectory. (b) Chromosome neighborhood search by monotonic decreasing learning ratio with non-dominated center in iteration n and n+ 1. These conditions are the sam¢ as VNS algorithm with dynamic neighborhood structure.