An intensified northern goshawk optimization algorithm for solving optimization problems

Abstract To enhance the performance of the Northern Goshawk Optimization (NGO) algorithm, which has limitations such as inadequate exploration, slow convergence rates, and susceptibility to local optima in later iterations, an Improved Northern Goshawk Optimization (INGO) algorithm has been developed. The INGO approach employs a good-point set strategy to obtain a high-quality initial population, generating evenly distributed solutions within the search domain. Incorporating an axial mapping differential variation method facilitates the creation of candidate solutions, broadening the scope for identifying optimal solutions, enhancing the algorithm’s ability to avoid local optima, and strengthening its global exploration capabilities. By implementing a Gaussian differential variation strategy, the diversity within the population is increased, which accelerates the convergence rate and improves the algorithm’s local search efficiency. For performance evaluation, the INGO algorithm was benchmarked against seven contemporary optimization algorithms using 65 test functions from the CEC2014, CEC2017, and CEC2022 suites. The outcomes indicate that INGO exhibits superior performance regarding convergence velocity, search precision, and stability when compared to the other algorithms.