教与学优化算法是一种新提出的群体智能优化算法，具备结构简单，无特定参数，容易实现等特点，越来越多的学者不断地对它进行改进并应用于更多的领域。

本文从基本的教与学优化算法出发，系统地研究了算法的原理、难点以及现有算法所存在的问题，结合其他群体智能优化算法的特性，分别提出了两种新的教与学优化算法并将改进算法应用到工程实例中。

首先，针对教与学优化算法初始种群质量较差，最优解值精度不佳等问题，利用Logistic-Tent混沌映射策略初始化种群，保证种群的多样性，在教师和学生阶段分别引入黄金正弦算法和基于莱维飞行与对数螺旋线的搜索策略优化个体的位置更新公式，增强并平衡算法的全局和局部收敛性能，提出了基于混合策略改进的教与学优化算法，进而选取十二组标准测试函数进行检验、将实验结果和迭代图像与原始的教与学优化算法以及其他群体智能优化算法进行对比分析，本文提出的新算法显著地提高了算法的稳定性和最优值的精度。

其次，针对教与学优化算法容易陷入局部最优值的问题，引入自适应授课因子和非线性动态惯性权重，并在此基础上增加了复习阶段，提出了基于自适应调整的教与学优化算法。在原有的十二个测试函数的基础上，增加六个固定维度多峰函数对其性能进行检验，将实验结果与原始的和先前改进的教与学优化算法的结果进行对比分析，本文提出的改进策略显著提高了算法的寻优精度和跳出局部最优的能力。

最后，基于自适应调整的教与学优化算法的提出主要是为了解决非线性约束优化问题，选取个三工程优化设计问题进行仿真实验，对比结果表明，该算法对工程实例优化问题的寻优具有更强的稳定性和更好的普适性，为求解该类问题提供了新的思路和方法。

Teaching-learning based optimization is a newly proposed swarm intelligence optimization algorithm, which has the characteristics of simple structure, no specific parameters, easy to implement, etc. More and more scholars continue to improve it and apply it in more fields.

This paper starts from the basic teaching-learning based optimization, systematically studies the principle of the algorithm, the difficulties and the existing problems of the algorithm, combined with the study of the characteristics of other optimization algorithms, respectively puts forward two new teaching-learning optimization algorithms and applies the improved algorithms to engineering examples.

Firstly, aiming at the problems such as poor quality of the initial population and poor accuracy of the optimal solution value of the teaching-learning based optimization algorithm, the Logistic-Tent chaotic mapping strategy was used to initialize the population and ensure the diversity of the population. The golden sine algorithm and the search strategy based on Levy flight and logarithm spiral were introduced to optimize the individual position updating formula in the teacher and student stage respectively. To enhance and balance the global and local convergence performance of the algorithm, mixed strategy based improved teaching-learning based optimization was proposed. Then 12 groups of standard test functions are selected for testing, and the experimental results and iterative images are compared with the original teaching and learning optimization algorithm. The new algorithm proposed in this paper significantly improves the stability of the algorithm and the accuracy of the optimal value.

Secondly, in order to solve the problem that the teaching-learning based optimization is easy to fall into the local optimal value, the adaptive teaching factor and nonlinear dynamic inertia weight are introduced, and the self-learning stage is added on this basis, and the self-adaptation adjustment of teaching-learning based optimization algorithm is proposed. On the basis of the original twelve test functions, six fixed dimension multimodal functions are added to test its performance. The experimental results are compared with those of the original and previous improved teaching-learning based optimization algorithm. The improved strategy proposed in this paper significantly improves the optimization accuracy and the ability to escape from the local optimal.

Finally, the self-adaptation adjustment of teaching-learning based optimization is proposed mainly to solve the nonlinear constraint optimization problem. Three engineering optimization design problems are selected for simulation. The comparison results show that the algorithm has stronger stability and better universality for the optimization of engineering case optimization problems, and provides a new idea and method for solving this kind of problem.