A Novel Multivariable MGM (1, m) Direct Prediction Model and Its Optimization

With regard to the traditional MGM (1, m) model having jumping error in solving process, an direct prediction (denoted as DMGM model) is proposed and its solution method put forward at first. Second, considering inherent time development trend of system behavior sequence ignored model, optimized by introducing a polynomial term, can be abbreviated TPDMGM m, φ ) model. Subsequently, it theoretically proved that id="M2"> achieve mutual transformation with adjusting parameter values. Finally, two case studies about predicting deformation foundation pit Henan’s vehicle ownership have been carried out validate effectiveness models. Meanwhile, Verhulst are established for comparison. Results show modeling performance four models from superior inferior ranked id="M3"> which on one hand testifies correctness defect analysis other verifies id="M4"> has advantages variables relation, restriction, characteristic.