An improved treble-level assisting optimization strategy to enhance algorithm search ability in heat exchanger network design

Abstract

Heat exchanger network synthesis is widely used for saving energy in the chemical engineering field. In the design stage, a multiple-level strategy is useful for assisting the optimization to guarantee the global search ability and protect the evolution process of the optimal solution. The authors have introduced a tri-level protection strategy based on the Random walk algorithm with compulsive evolution. However, the potential solutions in the middle level may still suffer from the operation of accepting the imperfect solution. The main deficiency of the tri-level protection strategy is that solutions are transferred to the elite level only from the protective level, ignoring the solutions with lower cost yield at the basic level. The strategy failed to fulfil the perfect protection and caused low computational efficiency. Therefore, an improved treble-level assisting optimization method is proposed in this paper, where the function of each level is redefined. Based on the advantages of the previous work, the efficiency is improved in the proposed strategy. The quantified data shows that the two cases are 2,782 $/yr and 7,134 $/yr lower than those reported in the literature, which verifies that the proposed strategy strengthens the global and local search abilities.