Salp Swarm Algorithm (SSA) is a newly developed meta-heuristic algorithm, which is mainly developed based on the swarming behavior of salps sailing and foraging in the ocean. An improved salp swarm-based optimizer is proposed in this study to overcome the potential shortcomings of original SSA, including being easily trapped in local or deceptive optima (LO) and its slow convergence rates in dealing with some high dimensional and multimodal landscapes. The designed variant is called CMSSA that combines two strategies, simultaneously. First, a chaotic exploitative mechanism (CEM) with “shrinking” mode is introduced into the basic SSA to improve the exploitative tendencies of the algorithm. Then, a combined mutation scheme is adapted to make full use of the strong intensification capabilities of Gaussian mutation and the strong exploratory leanings of Cauchy mutation. In addition, the embedded strategies can achieve a more stable equilibrium between the core searching patterns of the SSA, which are diversification and intensification. We thoroughly studied the optimization advantages of the improved CMSSA using several representative benchmark cases including unimodal, multimodal, and fixed-dimension multimodal functions, and three well-regarded engineering cases. The obtained experimental results, statistical tests, and comparative simulations indicate that the exploratory and exploitative proclivities of the SSA and its convergence patterns are vividly improved. The results indicate that the proposed CMSSA is a promising algorithm and shows superior efficacy compared to other algorithms.

This study aims to propose an efficient evaluation model for Sino-foreign cooperative education projects, which can offer a reasonable reference for universities to deepen reform and innovation of education and further enhance the level of international education. The core engine of the model is the kernel extreme learning machine (KELM) model integrated with orthogonal learning (OL) strategy optimization. The introduction of OL mechanism is to further strengthen the optimization capacity of the basic SCA, which is devoted to promoting the KELM model to select the optimal parameter combination and feature subset and further enhance the KELM evaluation capability of Sino-foreign cooperative education projects. For purpose of examining the performance of the proposed method, OLSCA is evaluated on 23 benchmark problems, comparison with eight other well-known methods. The experimental results have shown that the proposed OLSCA is prominently superior to existing methods on most functional problems. Meantime, OLSCA-KELM is comparison against other machine learning approaches in the evaluation of education projects of Sino-foreign cooperation. The simulation results illustrate that the presented OLSCA-KELM obtains better performance of classification and higher stability on all four indicators. Therefore, it is obvious that the presented OLSCA-KELM should be an effective solution for the evaluation of Sino-foreign cooperative education projects.