Abstract:
This A comprehensive job search from input space is crucial to securing the job search reliability of traditional job search algorithm-based methods since they involve enormous samples of fringes to be traversed from entire input domain. To this end, this paper aims to present two supervised algorithms for classification and reliability estimation namely Backpropagation Neural Network (BPNN), a supervised classifier algorithm and Levenberg-Marquardt Algorithm (LM), an optimizer that minimizes the error points are used respectively on breadth-first search (BFS) technique to solve the non-linear job scheduling problems for the reliability. Combined as a hybrid method (BPLM), the proposed algorithms tend to co-characterize a traversal efficient search domain through BPLM. While each algorithm enjoys special gain of its own, the BPNN can guarantee prescribed classifications and identifications from the search domain parameters during which the LM investigates large unseen BPNN processed input data for maximal reliability estimation through the use of mean square errors (MSE). In addition, the computational benefits of the proposed algorithms are scrutinized and their practicalities on efficiency and accuracy are grounded through comparative analysis with other published methods of engineering and science applications.