NOA ‐ RAC : An Enhanced Nutcracker Optimization Algorithm for Optimization Tasks

ABSTRACT Most real‐world problems are constrained continuous variable problems and discrete variable problems. In order to develop an algorithm that can solve these two types of problems in a balanced way, this paper proposes NOA‐RAC, an enhanced variant of the Nutcracker Optimization Algorithm (NOA). To address the limitations of NOA, including the exploration‐exploitation imbalance and tendency to fall into local optima in certain cases, three enhancement strategies were implemented. First, the implementation of a random subgroup strategy to better balance exploration‐exploitation trade‐offs. Second, the development of an adaptive fitness update mechanism that enhances population diversity. Finally, incorporation of a retractable transformable cruise strategy improves the algorithm's ability to jump out of local optima. A comprehensive experimental analysis, including effectiveness analysis of improvement strategies, qualitative analysis, non‐parametric statistical test, and so forth, was conducted to validate the results of the algorithmic improvements from multiple perspectives. NOA‐RAC was quantitatively compared with well‐known algorithms of various types proposed in recent years in three tests (CEC2017 benchmark suite, 30 engineering problems, and 12 feature selection problems). Experimental results demonstrate that NOA‐RAC exhibits strong competitiveness in solving both discrete‐variable and constrained continuous‐variable optimization problems, it serves as an effective tool for addressing real‐world optimization problems.