Browsing by Author "Tuncer, Adem"
Now showing 1 - 4 of 4
Results Per Page
Sort Options
Item Artificial Bee Colony Algorithm Based Hyper-Parameter Optimization for Convolutional Neural Networks(Yalova University, 2021-05-21) Özdemir, Koray; Özen, Yunus; Tuncer, AdemIn recent years, Deep Learning has become a field that researchers are particularly interested in. The Convolutional Neural Network (CNN) is a type of multi-layered artificial neural network mostly used in the analysis, recognition, and classification of images and videos. The performance of CNN models is usually based on custom model architectures, thus several hyper-parameter values in a CNN are manually selected mostly. However, different combinations of hyper-parameters and models need to be used to achieve better performance results. Determination of optimum values of hyper parameters is also an optimization problem. The metaheuristic optimization techniques are able to solve such problems. In this paper, we propose to use the Artificial Bee Colony (ABC) algorithm which is a meta-heuristic approach to automatically determine the optimum architecture of a CNN by means of hyper-parameters. The most effective hyperparameters in the performance of CNN models have been optimized, which are the number of layers, the number and size of filters, activation function, batch size, learning rate, optimizer, and dropout rate. We have evaluated our optimized architecture using the well-known Fashion-MNIST dataset. The results demonstrate that the proposed model using ABC improves the performance of a CNN model.Item Crew Scheduling Optimization with Artificial Bee Colony Algorithm(8th International Advanced Technologies Symposium (IATS’17), 2017-10-19) Sardoğan, Melike; Tuncer, AdemCrew scheduling is one of the most important optimization problems for airline companies. It is the scheduling of weekly or monthly work schedule under certain constraints, such as working hours and weekly permits. There are many studies using analytical and heuristic approaches in the literature in order to solve this problem. In studies using heuristic approaches, genetic algorithms are used frequently. In this study, an artificial bee colony algorithm, which is a heuristic method, is used instead of the approaches applied to the current problem. Weekly work schedules are optimized according to daily working hours and days off for crew scheduling under a number of different personnel. From the simulation results, it is clearly seen that the artificial bee colony algorithm produces successful results within reasonable time.Item Deep Reinforcement Learning Based Mobile Robot Navigation in Unknown Indoor Environments(Yalova University, 2021-05-21) Özdemir, Koray; Tuncer, Adem;The importance of autonomous robots has been increasing day by day with the development of technology. Difficulties in performing many tasks such as target recognition, navigation, and obstacle avoidance autonomously by mobile robots are problems that must be overcome. In recent years, the use of deep reinforcement learning algorithms in robot navigation has been increasing. One of the most important reasons why deep reinforcement learning is preferred over traditional algorithms is that robots can learn the environments by themselves without any prior knowledge or map in environments with obstacles. This study proposes a navigation system based on the dueling deep Q network algorithm, which is one of the deep reinforcement learning algorithms, for a mobile robot in an unknown environment to reach its target by avoiding obstacles. In the study, a 2D laser sensor and an RGBD camera has been used so that the mobile robot can detect and recognize the static and dynamic obstacles in front of itself, and its surroundings. Robot Operating System (ROS) and Gazebo simulator have been used to model the robot and environment. The experiment results show that the mobile robot can reach its targets by avoiding static and dynamic obstacles in unknown environments.Item Detection of Lane Changing Vehicles with Wavelet Transform and K-Nearest Neighbor Algorithm(Yalova University, 2021-05-21) Avcı, Yunus Emre; Koçal, Osman Hilmi; Tuncer, AdemTraffic management is getting more complicated due to the increasing urbanization rate day by day. Therefore, many models have been developed using smart transportation systems to overcome this problem. Lane changing, which is one of the important issues of smart transportation, is one of the basic driving behaviors that has a major impact on traffic efficiency, safety, and flow. Many various approaches have been presented in the literature for lane changing detection. In this study, a novel method for lane changing detection with a wavelet transform approach is presented. In the study, the pNEUMA dataset was used to evaluate the performance of the proposed method. In detecting lane changing, the azimuth angles of the vehicles were calculated using the WGS-84 coordinates in the dataset. Multi-level discrete wavelet transform, and lateral deviation were applied to the azimuth series of vehicles on a sample street in the dataset, and the data obtained were then classified with K-Nearest Neighbor Algorithm to determine whether there was a lane changing. In addition, the direction and time of the lane changing were determined by using the maximum amplitude obtained with wavelet transform methods. The proposed approach in the study achieved an average accuracy rate of 98%. Compared to other approaches, the proposed method has less computation complexity and therefore can find results more quickly.