Purpose Due to high efficiency and low carbon of natural gas, the consumption of natural gas is increasing rapidly, and the prediction of natural gas consumption has become the focus. The purpose of this paper is to employ a prediction technique by combining grey prediction model and trigonometric residual modification for predicting average per capita natural gas consumption of households in China. Design/methodology/approach The GM(1,1) model is utilised to obtain the tendency term, then the generalised trigonometric model is used to catch the periodic phenomenon from the residual data of GM(1,1) model for improving predicting accuracy. Findings The case verified the view of Xie and Liu: “When the value of a is less, DGM model and GM(1,1) model can substitute each other.” The combination of the GM(1,1) and the trigonometric residual modification technique can observably improve the predicting accuracy of average per capita natural gas consumption of households in China. The mean absolute percentage errors of GM(1,1) model, DGM(1,1), unbiased grey forecasting model, and TGM model in ex post testing stage (from 2013 to 2015) are 32.5510, 33.5985, 36.9980, and 5.2996 per cent, respectively. The TGM model is suitable for the prediction of average per capita natural gas consumption of households in China. Practical implications According to the historical data of average per capita natural gas consumption of households in China, the authors construct GM(1,1) model, DGM(1,1) model, unbiased grey forecasting model, and GM(1,1) model with trigonometric residual modification. The accuracy of TGM is the best. TGM helps to improve the accuracy of GM(1,1). Originality/value This paper gives a successful practical application of grey model GM(1,1) with the trigonometric residual modification, where the cyclic variations exist in the residual series. The case demonstrates the effectiveness of trigonometric grey prediction model, which is helpful to understand the modeling mechanism of trigonometric grey prediction model.

Purpose – The aggregation of information from a group of decision experts for developing collective opinion is the important question in practice. The purpose of this paper is to provide a group decision-making method via ordered weighted aggregation (OWA) operator and grey incidence analysis. Design/methodology/approach – In this study, OWA operator provides aggregation of attribute values to form an overall decision for each decision expert, and grey incidence model provides aggregation of decision experts’ evaluations to form overall score for each alternative. The example illustrates the procedure and practicability of the proposed model. Findings – A new thought for multiple attribute group decision-making problems is given. The proposed method produces an overall desirability score for each alternative. Practical implications – This is to obtain a more comprehensive and realistic solution to the given group decision-making problem. The proposed analysis method of group decision-making problems reveals vitality of grey systems theory. Originality/value – This paper combines OWA operator and grey incidence analysis to obtain a novel and effective method for group decision making. It is suitable for group decision-making problems in which the attribute weights are completely unknown, expert weights are completely unknown.

No abstract

Road traffic safety system is a grey system, which can be forecasted by the optimized Grey Verhulst model developed in this paper. The original Grey Verhulst model is optimized and improved from two aspects: the grey whitening derivative problem and the initial values. The new optimized model is established to forecast road traffic accidents, meanwhile the detail modeling steps is also given. Taking Chinese's road traffic accident statistics as an example, the optimized model is tested. The result shows that the improved Grey Verhulst model is with high prediction accuracy and low error, which provides a reference for the accurate prediction of road traffic accidents.

Purpose The purpose of this paper is to provide a new recursive GM (1,1) model based on forgetting factor and apply it to the modern weapon and equipment system. Design/methodology/approach In order to distinguish the contribution of new and old data to the grey prediction model with new information, the authors add forgetting factor to the objective function. The purpose of the above is to realize the dynamic weighting of new and old modeling data, and to gradually forget the old information. Second, the recursive estimation algorithm of grey prediction model parameters is given, and the new information is added in real time to improve the prediction accuracy of the model. Findings It is shown that the recursive GM (1,1) model based on forgetting factor can achieve both high effectiveness and high efficiency. Originality/value The paper succeeds in proposing a recursive GM (1,1) model based on forgetting factor, which has high accuracy. The model is applied to the field of modern weapon and equipment system and the result the model is better than the GM(1,1) model. The experimental results show the effectiveness and the efficiency of the prosed method.