This paper proposes a new metaheuristic algorithm with a search space reduction capability guided by simple formalism. The search population focuses partially on the inside the local search area while the rest explore globally, looking for better search areas. We call the new algorithm by YUKI Algoritm (YA) and employ it in a crack identification problem. With the aid of a set of measurements taken on the defected structure, we aim at identifying the crack parameters such as length and orientation. To this end, we use the so-called model reduction technique through Proper orthogonal Decomposition (POD) endorsed with Radial Basic Function (RBF), which helps in predicting (numerically) the measurement at new points (out of the set of sensors) via interpolation. This method is widely used in this context and was proven very effective computational-wise. In our study of the performance of YA, we deal with two cases; Firstly, in the case of the Elastostatic study. And secondly, in the case of dynamic analysis. We compare the performance of the suggested algorithm with the performance of well-known optimization methods, such as Teaching Learning Based Optimization (TLBO), Cuckoo Search (CS), and the Gray Wolf Optimizer (GWO). The results show that YA provides accurate and faster results compared to the mentioned algorithms.