No abstract

No abstract

Developing an early-warning system for air quality prediction and assessment of cities in China, Expert Systems With Applications (2017), Highlights  An early-warning system is developed for air quality.  Pollutant emission characteristics are analyzed using distribution functions.  Dynamic forecast intervals are constructed for addressing the uncertainty.  Air quality is evaluated by integrating fuzzy set theory and AHP.  The results show that the developed early-system is effective and reliable. Dear reviewer,Thank you very much for the positive and constructive comments on our manuscript entitled "Developing an early-warning system for air quality prediction and assessment: A case study in China". (No.: ESWA-D-17-00252R1). We have studied the comments and tried our best to revise the manuscript. Here below is our response to your comments and the revised parts in the revised manuscript are marked by highlight color,

The emerging complex circumstances caused by economy, technology, and government policy and the requirement of low-carbon development of power grid lead to many challenges in the power system coordination and operation. However, the real-time scheduling of electricity generation needs accurate modeling of electricity demand forecasting for a range of lead times. In order to better capture the non-linear and non-stationary characteristics and the seasonal cycles of future electricity demand data, a new concept of the integrated model is developed and successfully applied to research the forecast of electricity demand in this paper. The proposed model combines adaptive Fourier decomposition method, a new signal pre-processing technology, for extracting useful element from the original electricity demand series through filtering the noise factors. Considering the seasonal term existing in the decomposed series, it should be eliminated through the seasonal adjustment method, in which the seasonal indexes is calculated and should multiply the forecasts back to restore the final forecast. Besides, a newly proposed moth-flame optimization algorithm is used to ensure the suitable parameters of the least square support vector machine which can generate the forecasts. Finally, the case studies of Australia demonstrated the efficacy and feasibility of the proposed integrated model. Simultaneously, It can provide a better concept of modeling for electricity demand prediction over different forecasting horizons.

No abstract