Non-equidistance generalized accumulated grey forecast model and its application

Based on the generalized accumulated generating operation, this paper presents a new grey model, GAGM (1, 1), for the non-equidistant sequence with jump trend and multi-stage. According to the traditional GM (1, 1), the parameter identification of GAGM (1, 1) is introduced. The particle swarm optimization algorithm is used to solve the parameters. At last modeling with the production of gas well shows that the precision of GAGM(1, 1) is higher than other models. So GAGM(1, 1) generalizes grey model´s application range and has important theory significance.