基于多机制优化螺旋飞行特征的乌燕鸥算法

乌燕鸥优化算法(sooty tern optimization algorithm, STOA)是近年来出现的一种新兴的群智能优化算法，因其独特的螺旋式寻优方式和在实际工程问题中显著的优化效果而备受研究与关注。然而， STOA算法本身也存在着收敛速度较慢，搜索精度较低，并且容易陷入局部最优等缺点。因此，提出一种融合混沌映射、自适应惯性权重与高斯变异的多机制乌燕鸥优化算法(multi-mechanism sooty tern optimization algorithm incorporating chaotic mapping, adaptive inertia weight and gaussian mutation, GT-STOA)，以促进群体多样性并增强算法全局搜索和局部寻优的能力。同时，为验证算法寻优效果的显著性，基于12个具有不同特征的测试函数，选取九种典型的优秀元启发式算法进行对比验证。实验结果表明GT-STOA相较于其它九种算法具有更高的寻优精度和更快的收敛速度，并且易跳出局部最优找到全局最优解。此外，为了研究GT-STOA解决实际问题的能力，对压力容器设计问题进行优化求解，所得实验数据显示GT-STOA较传统STOA算法在求解精度上提升了42.54%。

Sooty Tern Algorithm Based on Multi-mechanism Optimized Spiral Flight Characteristics

The sooty tern optimization algorithm (STOA) is an emerging swarm intelligence optimization algorithm in recent years, which has attracted much research and attention because of its unique spiral optimization search method and the remarkable optimization effect in practical engineering problems. However, the STOA has the disadvantages of slower convergence, lower search accuracy, and easily fall into local optimum. Therefore, a multi-mechanism sooty tern optimization algorithm (GT-STOA), which combines chaotic mapping, adaptive inertia weight and gaussian mutation is proposed to promote population diversity and enhance the global search and local optimization capability of the algorithm. Meanwhile, to verify the significant optimization effect of the algorithm, nine typical and excellent meta-heuristic algorithms are selected for comparative validation on a total of 12 test functions with different features. The experimental results show that the GT-STOA has higher optimization accuracy and faster convergence speed than the other nine algorithms, and it is easy to jump out of the local optimum to find the global optimum solution. In addition, in order to study the ability of GT-STOA to solve practical problems, the optimization is conducted on the design of pressure vessel problem, and the experimental data obtained show that the GT-STOA improved the solving accuracy by 42.54% compared with the traditional STOA algorithm.