The Journal of Grey System ›› 2026, Vol. 38 ›› Issue (5): 11-21.

Previous Articles     Next Articles

Relationship Between Two Undetermined Parameters in a Trapezoidal Background Value Function

  

  1. 1. School of Traffic and Transportation, Shijiazhuang Tiedao University, Shijiazhuang, Hebei, 050043, P.R. China
    2. Department of Digital Equipment, Jiangsu Vocational College of Electronics and Information, Huai’an, Jiangsu, 223003, P.R. China
    3. School of Civil Engineering, Universiti Sains Malaysia, Engineering Campus, Nibong Tebal, Pulau Pinang, 14300, Malaysia
  • Online:2026-11-16 Published:2026-11-16

Abstract: Accurate background-value construction is critical to GM(1,1) performance. This study improves Zhang’s trapezoidal background-value function by using quadratic Newton interpolation to approximate the first-order accumulated sequence and derive its theoretical integral background value. An equivalent mapping between the integral and trapezoidal background values is established, yielding an analytical coupling relationship between the point-position and slope parameters. This coupling eliminates independent slope-parameter optimization and reduces parameter-identification dimensionality. Truncation-error characteristics and error propagation to model-parameter estimation are analyzed, and the feasible range of the point-position parameter is geometrically extended. Two improved models are developed: Model I with a fixed point-position parameter and Model II with a variable-weight mechanism. Repeated comparisons identify the grey wolf optimizer (GWO) as the most stable algorithm. Case studies with fitting-sequence lengths of 6, 7, and 8 show that the proposed models generally outperform nine benchmark grey models in fitting and forecasting accuracy, demonstrating improved efficiency, robustness, and predictive performance.

Key words: GM(1,1) model, Background value, Quadratic Newton interpolation, Parameter coupling, Variable-weight mechanism, Grey wolf optimizer