Abstract
In the real world, as the complexity of optimization problems continues toincrease, there is an urgent need to research more efficient optimizationmethods. Current optimization algorithms excel in solving problems with a fixednumber of dimensions. However, their efficiency in searching dynamicmulti-dimensional spaces is unsatisfactory. In response to the challenge ofcross-dimensional search in multi-dimensional spaces with varying numbers ofdimensions, this study proposes a new optimization algorithm-Dynamic DimensionWrapping (DDW) algorithm. Firstly, by utilizing the Dynamic Time Warping (DTW)algorithm and Euclidean distance, a mapping relationship between different timeseries across dimensions is established, thus creating a fitness functionsuitable for dimensionally dynamic multi-dimensional space. Additionally, DDWintroduces a novel, more efficient cross-dimensional search mechanism fordynamic multidimensional spaces. Finally, through comparative tests with 31optimization algorithms in dynamic multidimensional space search, the resultsdemonstrate that DDW exhibits outstanding search efficiency and provides searchresults closest to the actual optimal solution.