Ensemble Shuffled Population Algorithm for multi-objective thermal design optimization of a plate frame heat exchanger operated with Al2O3/water nanofluid

This study proposes a brand new optimization algorithm entitled Ensemble Shuffled Population Algorithm for solving multidimensional optimization problems. The proposed algorithm adopts the perturbation equations of the Crow Search and Differential Search algorithms with useful modifications on them and aims to maintain a reasonable balance between the intensification and diversification phases of the algorithm. A batch of 22 benchmark problems consisting of unimodal and multimodal unconstrained optimization test functions are applied using this algorithm to assess its performance on multi dimensional problems. Statistical results obtained from the proposed Ensemble Shuffled Population Algorithm are compared to those found by eleven well known metaheuristic optimizers. The comparison results show that the Ensemble Shuffled Population Algorithm outperforms the compared optimizers with regards to solution accuracy and convergence speed. After that, the proposed algorithm is applied on a multi objective optimization of a plate frame heat exchanger operated with Al2O3 nanofluid. The optimization results show that utilizing nanoparticles instead of base fluid not only increases the overall heat transfer coefficient rates but also entails a huge decline in total cost values. A Pareto frontier is constructed for these two conflicting objectives to select the final optimum solution from the set of non-dominated solutions by virtue of three famous decision making methods of LINMAP, TOPSIS, and Shannon's entropy theory. Then, sensitivity analysis is performed to observe the variational effects of the design variables on the optimization objectives.