Manar Hosny

A mutation-based genetic algorithm for room and proctor assignment in examination scheduling

Examination scheduling is a very important task that has to be done in all academic institutions periodically. Formulating exam schedules manually requires immense time and effort, due to the presence of a large number of conflicting constraints that must be satisfied. In this study, we tackle the examination scheduling problem that is specific to the female section in our college, and particularly to the Masters program. Due to cultural restrictions, different room types may be needed to schedule exams, if the instructor is of a different gender than the students. In addition, proctors should be assigned to supervise these exams. We propose a Genetic Algorithm (GA) approach to solve the problem. Our approach follows the classical GA framework but without the crossover operator. We consider mutation as the main genetic operator during the evolutionary process, in order to avoid disruption of constraints and maintain the feasibility of solutions as much as possible. For our examination scheduling problem, two optimization phases have been developed. In the first phase, we find the best room assignment, in terms of room type and the appropriate number of seats for each exam. While in the second phase, the exams will be assigned to proctors for supervision. Each of these phases has a different set of hard constraints that have to be satisfied in the solution. In addition, there are also soft constraints, which should be optimized to improve the quality of the solution. The experimental results indicate the efficiency of the algorithm in handling the constraints that are specific to this examination scheduling problem.

Human Factors in the Design of BCI-Controlled Wheelchairs

In this paper, we synthesize research on the type of cognitive commands that have been examined for controlling Brain Computer Interface (BCI) wheelchairs and the human factors that have been reported for the selection of different protocols of BCI commands for an individual user. Moreover, we investigate how different researchers have considered the necessity of sustained movement from a single thought/command, having an emergency stop, and the commands necessary for assisting users with a particular disability. We then highlight how these human factors and ergonomics’ considerations were applied in the design and development of an EEG-controlled motorized wheelchair, aiming to emphasize users’ requirements for people with severe physical disabilities. In this case study, we propose a brain controlled wheelchair navigation system that can help the user travel to a desired destination, without having to personally drive the wheelchair and frequently change the movement directions along the path to the destination. The user can choose the desired destination from a map of the environment, using his/her brain signals only. The user can navigate through the map using BCI cognitivecommands. The system processes the brain signals, determines the required destination on the map, and constructs an optimized movement path from the source to the intended destination. To construct an obstacle-free path with the shortest possible distance and minimum number of turns, a path planning optimization problem is solved using a simple Simulated Annealing (SA) algorithm. The resulting optimized path will be translated into movement directions that are sent to the microcontroller to move the wheelchair to the desired destination.

A Multimodal Adaptive Genetic Clustering Algorithm

Clustering is widely used in a variety of fields to find structures among data and extract useful knowledge. Recently, there has been an emergent need for robust and efficient techniques that can manage the exploding volume of data available in the World Wide Web or gathered from devices and sensors. However, clustering such data is challenging, due to the multimodal nature of this information. In this work, we introduce a novel Multimodal Adaptive Genetic Clustering (MAGC) algorithm that clusters information based on multiple features. Our approach adds feature weights as an extension to the chromosome, which represents a clustering solution, such that feature weights are also evolved and optimized alongside the original clustering solution. The number of clusters is also adaptive and is optimized during the search.

An optimized single-finger Arabic keyboard layout

Cellular phones and other hand-held devices are now extensively used to write emails, notes and long texts. However, the arrangement of keys in the current keyboards is not optimized to facilitate rapid and ergonomic typing. In this paper, we aim to optimize the Arabic keyboard layout for applications that predominantly use a single pointer. The single-finger keyboard layout problem can be modeled in terms of the Quadratic Assignment Problem (QAP), which can be solved using metaheuristic algorithms. To adapt the problem to the requirements of optimizing the single-finger keyboard, we used three measures in our objective function: the distance between pairs of letters, a weight for each row in the keyboard, and the hit direction of the finger. A Genetic Algorithm (GA) approach with two crossover types (2-point and modified uniform crossovers) and two different mutation operators (swap and insertion) was developed and thoroughly tested. The performance of the Genetic Algorithm was also compared against a Simulated Annealing (SA) algorithm using the same objective function. Moreover, we developed a Memetic Algorithm combining GA and SA to maximize the chances of obtaining good solutions. We compared our resulting optimized keyboard layouts with different existing and proposed layouts. The comparison results show that our keyboard layouts are more efficient, in terms of the optimization criteria considered, than the tested layouts. Finally, the performance of our keyboards was tested by virtually estimating the speed of typing. Our keyboards also outperform other layouts in terms of the measured typing speed. The details of the algorithms and the experimental results are reported in this paper.

Recognition of Affective States via Electroencephalogram Analysis and Classification

Understanding and reacting to the affective state of users is increasingly becoming important in the field of human–computer interaction (HCI) research and practice. Recent developments in brain–computer interface (BCI) technology has facilitated improved accuracy in human emotion detection and classification. In this paper, we investigate the possibility of using electroencephalogram (EEG) for the detection of four affective states based on a dimensional model (valence and arousal) of emotions. We conduct rigorous offline analysis for investigating the deep neural network (DNN) classification method in emotion detection. We also compare our classification performance with a random forest (RF) classifier and support vector machine (SVM). The data analysis results revealed that the proposed DNN-based classifier method outperformed the methods based on the SVM and RF classifiers.

Measuring and monitoring emotional changes in children who stutter

The assessment of clients with speech disorders presents challenges for speech-language pathologists. For example, having a reliable way of measuring the severity of the case, determining which remedial program is aligned with a patients needs, and measuring of treatment processes. There is potential for brain-computer interface (BCI) applications to enhance speech therapy sessions by providing objective insights and real-time visualization of brain activity during the sessions. This paper presents a study on emotional state detection during speech pathology. The goal of this study is to investigate affective-motivational brain responses to stimuli in children who stutter. To this end, we conducted an experiment that involved recording frontal electroencephalography (EEG) activity from fifteen children with stuttering whilst they looked at visual stimuli. The contribution of our study is to provide a comprehensive background and a framework for emotional state detection experiments as assessment and monitoring tool in speech pathology. It mainly discusses the feasibility and potential benefits of applying EEG-based emotion detection in speech-language therapy contexts of use. The findings of our research indicate that emotional recognition using non-invasive EEG-based BCI system is sufficient to differentiate between affective states of individuals in treatment contexts.

Classification of Human Emotions from Electroencephalogram (EEG) Signal using Deep Neural Network

Estimation of human emotions from Electroencephalogram (EEG) signals plays a vital role in developing robust Brain-Computer Interface (BCI) systems. In our research, we used Deep Neural Network (DNN) to address EEG-based emotion recognition. This was motivated by the recent advances in accuracy and efficiency from applying deep learning techniques in pattern recognition and classification applications. We adapted DNN to identify human emotions of a given EEG signal (DEAP dataset) from power spectral density (PSD) and frontal asymmetry features. The proposed approach is compared to state-of-the-art emotion detection systems on the same dataset. Results show how EEG based emotion recognition can greatly benefit from using DNNs, especially when a large amount of training data is available.

BeamGA Median: A Hybrid Heuristic Search Framework

BeamGA is a general hybrid heuristic framework that can be used to solve the median problem in comparative genomics, where any distance function can be used. It starts with a heuristic search approach (local beam search) in order to generate a number of solutions. Then a Genetic Algorithm (GA) is applied to refine the solutions. It considers true biological evolution scenarios by applying the concept of common intervals during the GA optimization process.

A Genetic Algorithm Approach for Optimizing a Single-Finger Arabic Keyboard Layout

The use of cellphones and handheld devices in our daily activities is not limited to making calls or writing short text messages. The added features of wireless technology and related applications made it possible to write emails, notes and long text. Nevertheless, the currently used keyboards in portable devices are not optimized for such use, in terms of rapid and ergonomic typing. In this research, we aim to optimize the design of the Arabic keyboard layout for applications that predominantly use a single pointer, such as those used in portable devices. The main objective is to find the best single-finger Arabic keyboard layout that allows users of portable devices to write text and carry out written conversations for a long time with comfort, ease, and speed. Since the single-finger keyboard layout problem can be modeled in terms of the famous Quadratic Assignment Problem (QAP), which is known to be NP-hard, heuristics and meta-heuristics are recommended for solving such problem. To adapt the problem to the requirements of optimizing the single-finger Arabic keyboard, we added two measures to the classical—distance based—objective function of the QAP, which are: the keyboard row weight and the hit direction of the finger. A Genetic Algorithm (GA) approach with two different crossover types (two-point and modified uniform crossovers), and three different mutation operators (swap, insertion, and Simulated Annealing (SA)) was developed and thoroughly tested. The experimental results demonstrated that the simple swap mutation produced better results than the other mutations, with both crossover types. Moreover, experimental testing has shown that the added measures in the objective function had a positive effect, in terms of improving the typing speed, when compared to the original QAP objective function. Finally, comparing the resulting optimized keyboard layout with other existing keyboards showed that our keyboard layout is favorable, in terms of the optimization criteria considered in this research, than the other layouts tested.

Indoor Wheelchair Navigation for the Visually Impaired

Visually impaired (VI) people face many daily life challenges. It is often difficult for them to recognize where they are, and they may feel disoriented or completely isolated. Moreover, people who have other motor disabilities besides visual impairment face even more difficulty. For example, when using a wheelchair, they usually need a personal assistance to help them navigate to their destination, since they cannot use a cane or other manual assistive devices. To help these people, we aim to develop an indoor wheelchair navigation system. The system is divided into three components: a positioning system, a navigation system and a VI system interface. In the current research, we focus on the design of the navigation system, which will build the optimal path for a VI user on an electric wheelchair. To navigate, the user chooses a destination and their route preferences. A specially designed algorithm will construct an optimized path to the destination, by processing a map of the environment and the route preferences. The algorithm will take into account particular route features customized to the needs of the VI, such as: being free of obstacles, having few turns, being close to walls, and accommodating clues and landmarks. Our navigation system uses the A* shortest path algorithm for path construction, after adapting the objective function to take into account multiple criteria that fit the requirements of a visually impaired. The output of the algorithm is the moving directions, which will be fed back to the wheelchair as commands to direct its movement to the desired destination.