Seasonal grey Bernoulli model with cumulative effects based on the improved marine predators algorithm and its application

Due to demand variations, supply adjustments and price fluctuations, natural gas production (NGP) has a nonlinear complex growth trend with seasonal cumulative effects. We aim to construct a seasonal forecast model that identifies cumulative effects and reflects nonlinear complexity. First, a novel seasonal weighted fractional nonlinear grey Bernoulli model (SWFNGBM(1,1|sin)) is proposed. The seasonal weighted fractional accumulation generation operator (SWFAGO) is proposed to weaken the seasonal characteristics and complex fluctuations. The sinusoidal and linear terms are introduced to characterize the cumulative effects of the generated sequence. Then, model nonlinear parameters are optimized using an improved marine predators algorithm (LTEMPA). Logistic-Tent system (LTS) initialization and elite oppositional-based learning (EOBL) are used to improve population diversity and the quality of optimal solution. An improved control parameter is proposed to enhance the search capability in later stages. The performance of LTEMPA and SWFNGBM(1,1|sin) is validated by CEC 2017 benchmark test suite and case experiments. Finally, the quarterly NGP in China from 2024 to 2027 is predicted, and it will rise in the next four years. The growth trend is expected to slow down in 2027. The reasonableness of forecast results is analyzed and some future suggestions are given.