559 The Assessment of Applying Chaos Theory for Daily Traffic Estimation Abbas Mahmoudabadi Department of Industrial Engineering Payam-e-Noor University Tehran, Iran Saeed Andalibi High Institute of Education and Research in Planning and Management Tehran, Iran Abstract Road traffic volumes in intercity roads are generally estimated using probability functions, statistical techniques of regression or meta-heuristic approaches such as artificial neural networks. Although, road traffic volumes may be dependent upon input variables of road geometrical design, weather conditions, weekend, national holiday and so on, but traffic volumes are also estimated by using pattern recognition techniques. The main question from the researchers, here, is to check the using chaotic pattern of daily traffic volume and the performance of chaos theory to estimate daily traffic volume. In this research work, the existing of chaotic behavior in daily traffic volume in intercity roads has been examined while the performance of chaos theory is discussed compared to probability functions. The ratio between the minimum and maximum of daily traffic volume is considered as chaos factor. Data, gathered through two installed automatic traffic counters over a year, have been used in analytical process. Results show that daily traffic volumes have chaotic behavior with defined twenty-four hour time span, but applying the principle of chaos theory is not appropriate method for estimating daily traffic volumes comparing to a good fitness probability function. Keywords Chaos Theory, Traffic Volume Estimation, Probability Function, Pattern Recognition 1. Introduction Estimating traffic volume is one of the most important concerns for traffic managers, in particular for improving enforcement and planning traffic management. Collecting data and predicting traffic patterns are usually performed for pavement design, fuel-tax revenue projection and highway planning. But monitoring activities, necessary for accurate annual average daily traffic (AADT) estimates, are expensive in terms of costs and personnel (Ricardo R., 2012). Traffic estimation is not only a main issue for tactical purposes of transportation such as enforcement (Ashok, 2002), but also it is highly related to road accidents' frequency and their severities (Ardekani, 1995). On the other hand, transport responsibilities mainly focus on predicting road traffic volumes in order to make road safety plans over the network regarding to improving enforcement and implementing road safety measures. In the era of using intelligent transport systems, established in all over the world to improve the efficiency of transportation, traffic estimation has become more and more critical issue in road transport management (Caggiani, 2012). Estimating road traffic volumes has become more important while it is considered to determine origin- destination matrix as a basic stage of transportation planning (Caggiani, 2012). There are many models and methodologies for estimating traffic volumes in the literature. The main techniques including Regression Analysis (Faghri & Hua, 1995), Ward’s Minimum-Variance method of clustering (Sharma & Werner, 1981), Clustering-based methods (Zhao, et al., 2004), and some heuristic methods of Genetic Algorithms (Lingras, 2001) and Artificial Neural Networks (Faghri & Hua, 1995, Lingras, 2001, Mahmoudabadi & Fakharian, 2010) have been implemented for road traffic estimation. Simulation techniques (Juran et al., 2009) and data mining (Gecchele et al., 2011) which are mainly developed through using the probability distribution functions are also applied for traffic estimation. Variation in day-to-day traffic volumes is a very important factor in the process of Proceedings of the 2014 International Conference on Industrial Engineering and Operations Management Bali, Indonesia, January 7 – 9, 2014