Scholarlink

<p>To address the issues of low search efficiency, poor convergence accuracy, and susceptibility to local optima in the sand cat swarm optimization (SCSO) algorithm, this paper proposes a modified sand cat swarm optimization algorithm incorporating gold exploration strategy and sparrow alert mechanism, termed the GE-SA ISCSO algorithm. Firstly, Bernoulli chaotic mapping is employed to generate a more uniformly distributed initial population. Secondly, the gold exploration strategy is embedded into the position update formula of sand cats, providing more candidate search directions for locating prey. After the SCSO algorithm search phase is completed, a sparrow alert mechanism is further introduced to accelerate convergence and enhance the ability to escape local optima. To verify the effectiveness and superiority of GE-SA ISCSO, comparative tests are conducted on 21 classical benchmark functions and the CEC2022 test set function against several metaheuristic optimization algorithms. The performance of GE-SA ISCSO is evaluated using the Wilcoxon rank sum test, box plot analysis, and five qualitative methods. Test results show that GE-SA ISCSO achieves higher convergence accuracy, faster convergence speed, and a stronger capability to avoid local optimal solutions, demonstrating outstanding performance in solving complex numerical optimization problems. In addition, six engineering design problems and a photovoltaic power forecasting task are introduced to further assess the comprehensive performance of the proposed algorithm. For all six engineering design problems, GE-SA ISCSO attains the best results among all comparative algorithms, exhibiting excellent optimization capability. In particular, compared with the original the SCSO algorithm, the best objective values obtained by GE-SA ISCSO are reduced by 15.42%, 18.41 %, 0.067%, 29.03%, 0.157%, and 99.63% on the six problems, respectively, fully demonstrating its applicability and superiority in handling real-world engineering optimization problems with complex constraints. When applied to photovoltaic power forecasting, the GE-SA ISCSO-LSTM model significantly improves prediction accuracy. Under sunny conditions, compared with the SCSO-LSTM model, the mean squared error (MSE), mean absolute error (MAE), root mean squared error (RMSE), mean absolute percentage error (MAPE), and relative root mean squared error (RRMSE) are reduced by 97.16%, 84.22%, 83.13%, 64.39%, and 83.14%, respectively; under rainy conditions, these errors are reduced by 84.64%, 63.86%, 60.81%, 82.76%, and 60.81%, respectively. These findings indicate that the proposed the GE-SA ISCSO algorithm is of significant practical value for reducing costs and improving resource utilization in engineering optimization and renewable energy applications.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A modified sand cat swarm optimization algorithm incorporating gold exploration strategy and sparrow alert mechanism

  • 期刊论文

  • 发表日期：

    2026年4月30日

查看全文
  • Lingzhi Wang,
  • Jiarui Zhao,
  • Chenyang Li,
  • Wanjun Li,
  • Cheng Li,
  • Jie Li,
  • Ling Zhao

摘要

To address the issues of low search efficiency, poor convergence accuracy, and susceptibility to local optima in the sand cat swarm optimization (SCSO) algorithm, this paper proposes a modified sand cat swarm optimization algorithm incorporating gold exploration strategy and sparrow alert mechanism, termed the GE-SA ISCSO algorithm. Firstly, Bernoulli chaotic mapping is employed to generate a more uniformly distributed initial population. Secondly, the gold exploration strategy is embedded into the position update formula of sand cats, providing more candidate search directions for locating prey. After the SCSO algorithm search phase is completed, a sparrow alert mechanism is further introduced to accelerate convergence and enhance the ability to escape local optima. To verify the effectiveness and superiority of GE-SA ISCSO, comparative tests are conducted on 21 classical benchmark functions and the CEC2022 test set function against several metaheuristic optimization algorithms. The performance of GE-SA ISCSO is evaluated using the Wilcoxon rank sum test, box plot analysis, and five qualitative methods. Test results show that GE-SA ISCSO achieves higher convergence accuracy, faster convergence speed, and a stronger capability to avoid local optimal solutions, demonstrating outstanding performance in solving complex numerical optimization problems. In addition, six engineering design problems and a photovoltaic power forecasting task are introduced to further assess the comprehensive performance of the proposed algorithm. For all six engineering design problems, GE-SA ISCSO attains the best results among all comparative algorithms, exhibiting excellent optimization capability. In particular, compared with the original the SCSO algorithm, the best objective values obtained by GE-SA ISCSO are reduced by 15.42%, 18.41 %, 0.067%, 29.03%, 0.157%, and 99.63% on the six problems, respectively, fully demonstrating its applicability and superiority in handling real-world engineering optimization problems with complex constraints. When applied to photovoltaic power forecasting, the GE-SA ISCSO-LSTM model significantly improves prediction accuracy. Under sunny conditions, compared with the SCSO-LSTM model, the mean squared error (MSE), mean absolute error (MAE), root mean squared error (RMSE), mean absolute percentage error (MAPE), and relative root mean squared error (RRMSE) are reduced by 97.16%, 84.22%, 83.13%, 64.39%, and 83.14%, respectively; under rainy conditions, these errors are reduced by 84.64%, 63.86%, 60.81%, 82.76%, and 60.81%, respectively. These findings indicate that the proposed the GE-SA ISCSO algorithm is of significant practical value for reducing costs and improving resource utilization in engineering optimization and renewable energy applications.