Multi-gene genetic programming based predictive models for municipal solid waste gasification in a fluidized bed gasifier

• Genetic programming is used to predict the performance of fluidized bed gasifier.
• The performance of the MGGP models is compared with the single-gene GP model.
• Comparisons of complexity and accuracy of GP prediction have been reported.
• The MGGP approach gives better results on both training and validation data.
• The data-driven GP modelling is useful for prediction with analytical expressions.

A multi-gene genetic programming technique is proposed as a new method to predict syngas yield production and the lower heating value for municipal solid waste gasification in a fluidized bed gasifier. The study shows that the predicted outputs of the municipal solid waste gasification process are in good agreement with the experimental dataset and also generalise well to validation (untrained) data. Published experimental datasets are used for model training and validation purposes. The results show the effectiveness of the genetic programming technique for solving complex nonlinear regression problems. The multi-gene genetic programming are also compared with a single-gene genetic programming model to show the relative merits and demerits of the technique. This study demonstrates that the genetic programming based data-driven modelling strategy can be a good candidate for developing models for other types of fuels as well.