Energy and Buildings 139 (2017) 254–262 Contents lists available at ScienceDirect Energy and Buildings j ourna l ho me pa g e: www.elsevier.com/locate/enbuild An investigation for predicting the effect of green roof utilization on temperature decreasing over the roof surface with Gene Expression Programming Tahir Ayata a,∗ , Dogan Erdemir b , Omer Tayfun Ozkan b a Erciyes University, KMY, 38039 Kayseri, Turkey b Erciyes University, Department of Mechanical Engineering, 38039 Kayseri, Turkey a r t i c l e i n f o Article history: Received 25 March 2016 Received in revised form 21 November 2016 Accepted 7 January 2017 Available online 9 January 2017 Keywords: Green roof Gene Expression Programming Temperature decreasing Comfort condition a b s t r a c t This study presents that the effect of green roof usage on the temperature decreasing for Turkey’s cities has been investigated with Gene Expression Programming (GEP). Training data for GEP model has been taken from summer climatic data of nine different cities around the world, which have different climate regions. Thus, wide range of climatic conditions have been considered in GEP modeling. GEP model has been used for predicting the temperature decreasing over a green roof in Turkey cities. GEP model has given sufficient result with 2.2894% RMSE and 94.01% R 2 . When the effect of green roof usage in Turkey cities is compared to World cities, it is found that temperature decreasing value for Turkey’s cities and World cities has shown same trend. The maximum temperature increasing is 2.91 ◦ C in Kars at 08:00 a.m. and the maximum temperature decreasing is 29.61 ◦ C in Izmir at 06:00 p.m. © 2017 Elsevier B.V. All rights reserved. 1. Introduction The green roof systems have started to be used commercially. It is possible to contribute to the comfort requirements of the building by increasing the insulation effect and heat gain from building by installing a green roof system on the building roof [1]. In classic rural buildings, green roof systems have been used in a natural way from the past [2]. In improving home comfort conditions, thermal insu- lation effect and conditioning properties of green roof systems are now being used widely [3]. In this case, it is important to evaluate the performance of green roofs according to the climatic conditions of the cities. In many studies, standard concrete roof and green roof were compared in terms of heat losses, heat gains and temperature decreases. These comparisons give different results depending on the geographical zone and climate conditions. In one of these studies, experimental and theoretical temperature analyses were performed and created a heat transfer model of a standard green roof system [4]. In another study, the average temperature varia- tion was observed between a standard roof and a green roof within three week-period experimental measurements, in July [5]. Inves- tigation of the result of the green roof system for different climatic ∗ Corresponding author. E-mail address: tahirayata@yahoo.com (T. Ayata). conditions is also important. Studies for the different geographical and climatic regions show that the surface temperature decrease of the green roofs is significant in all climatic conditions [6]. A limited quantity of studies have observed green roof effects in win- ter. Green roofs can act as insulation during cold weather, although properly applied current insulation under a green roof will increase the insulating value. In one such study it is predicted that soil mois- ture, a thermal mass contributor, may be the essential factor in the insulating capacity of green roof systems in winter [7]. As known, the lack of heat gain of a roof in summer and the absence of heat loss in winter is desirable. For this purpose, green roof system has been investigated experimentally in terms of heat loss and gain effects on buildings [8,9]. As a result of these studies, the green roof yields heat gain instead of heat loss in winter, while it supplies reduction on heat gain in summer. Green vegetation stands out with positive characteristics in every climatic condition [10]. With the increasing number of experimental studies in this field, it is now possible to reach large number experimental data of the green roofs and to make calculations with different theoretical and numerical methods [11–14]. The passive cooling potential of green roof is predicted with the obtained experimental results used in ANN learning data [15,16]. The traditional roof and the green roof comparison has been inves- tigated for some European cities by using numerical methods [17]. Also, studies comparing ANN and other convergence methods are also available [18]. Besides, there are many studies using genetic http://dx.doi.org/10.1016/j.enbuild.2017.01.014 0378-7788/© 2017 Elsevier B.V. All rights reserved.