“…1, the two-dimensional normalized criterion space of each variant of flow path design is characterized by the corresponding point, whose distance to the center of the coordinates is proportional to the value ‖ ( , )‖. In essence, this form of convolution (2) is similar to the convolution of the ideal point method [8], which is in turn close to Germeyer's convolution [9]. As shown by Karpenko et al [8], these convolutions, make it possible to find optimal solutions, both for convex and for non-convex Pareto sets.…”