Artificial intelligence methods in breakwater damage ratio estimation O. Yagci * , D.E. Mercan, H.K. Cigizoglu, M.S. Kabdasli Division of Hydraulics, Civil Engineering Faculty, Istanbul Technical University, 34469 Maslak, Istanbul, Turkey Received 22 September 2004; accepted 8 March 2005 Available online 20 June 2005 Abstract The anticipation of damage ratio with an acceptable accuracy is a vital issue in breakwater design. The presented study covers the employment of three different artificial neural network methods and a fuzzy model for this problem. Inputs like mean wave period, wave steepness, significant wave height and the breakwater slope are used as input to estimate the corresponding damage ratio value. All artificial neural network methods and fuzzy logic model provided quite close estimations for the experimental values. The testing stage results were significantly superior to the conventional multi- linear regression method in terms of the selected performance criteria. q 2005 Elsevier Ltd. All rights reserved. Keywords: Armor unit; Artificial intelligence; Breakwater; Damage ratio; Fuzzy logic; Neural network; Static stability 1. Introduction Breakwater stability analysis has long been attracted the interest of coastal engineering researchers. The design of armor layer units on breakwater is one of the major problems for coastal engineers. The type, weight and placement technique of breakwaters 0 armor layer units are designed considering anticipated damage ratio, which will occur under the estimated wave climate conditions. Therefore, the anticipation of damage ratio with an Ocean Engineering 32 (2005) 2088–2106 www.elsevier.com/locate/oceaneng 0029-8018/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.oceaneng.2005.03.004 * Corresponding author. Tel.: C90 212 2856011; fax: C90 212 2856587. E-mail address: oyagci@ins.itu.edu.tr (O. Yagci).