Ahmed Fouad Ali

Tabu search with multi-level neighborhood structures for high dimensional problems

Metaheuristics have been successfully applied to solve different types of numerical and combinatorial optimization problems. However, they often lose their effectiveness and advantages when applied to large and complex problems. Moreover, the contributions of metaheuristics that deal with high dimensional problems are still very limited compared with low and middle dimensional problems. In this paper, Tabu Search algorithm based on variable partitioning is proposed for solving high dimensional problems. Specifically, multi-level neighborhood structures are constructed by partitioning the variables into small groups. Some of these groups are selected and the neighborhood of their variables are explored. The computational results shown later indicate that exploring the neighborhood of all variables at the same time, even for structured neighborhood, can badly effect the progress of the search. However, exploring the neighborhood gradually through smaller number of variables can give better results. The variable partitioning mechanism used in the proposed method can allow the search process to explore the region around the current iterate solution more precisely. Actually, this partitioning mechanism works as dimensional reduction mechanism. For high dimensional problems, extensive computational studies are carried out to evaluate the performance of newly proposed algorithm on large number of benchmark functions. The results show that the proposed method is promising and produces high quality solutions within low computational costs.

Genetic algorithm with population partitioning and space reduction for high dimensional problems

In this paper, we modify genetic algorithm (GA) with new strategies of population partitioning and space reduction for high dimensional problems. The proposed method is called GA with Matrix-Coding Partitioning (GAMCP). In the GAMCP method, a population of chromosomes is coded in a one big matrix. This matrix is partitioned into several sub-matrices every generation, and GAMCP applies the genetic operations on the partitioned sub-matrices. Moreover, the Gene Matrix (GM) [5], [6] termination criteria are modified and applied in the GAMCP method in order to equip the search process with a self-check to judge how much exploration has been done and to maintain the population diversity. The computational experiments show the efficiency of the new elements proposed in the GAMCP method.

Minimizing molecular potential energy function using genetic Nelder-Mead algorithm

This paper presents a new algorithm for minimizing the molecular potential energy function. The new algorithm combines a global search genetic algorithm with a local search Nelder-Mead algorithm in order to search for the global minimum of molecular potential energy function. The minimization of molecular potential energy function problem is very challenging, since the number of local minima grows exponentially with the molecular size. The new algorithm is called GNMA (Genetic Nelder-Mead Algorithm). Such hybridization enhances the power of the search technique by combining the wide exploration capabilities of Genetic Algorithm (GA) and the deep exploitation capabilities of Nelder-Mead algorithm. The proposed algorithm can reach the global or near-global optimum for the molecular potential energy function with up to 200 degrees of freedom. The performance of the proposed algorithm has been compared with other 9 existing methods from the literature. The numerical results show that the proposed algorithm is promising and produce high quality solutions with low computational costs.

Nature Inspired Optimization Algorithms for CT Liver Segmentation

Nature inspired optimization algorithms have gained popularity in the last two decades due to their efficiency and flexibility when they applied to solve global optimization problems. These algorithms are inspired from the biological behavior by swarms of birds, fish and bees. In this chapter, we give an overview of some of nature inspired optimization algorithms such as Artificial Bee Colony (ABC), Cuckoo Search (CS), Social Spider Optimization (SSO) and Grey Wolf Optimization (GWO). Also, we present the usage of ABC and GWO algorithms for CT liver segmentation. The experimental results of the two selected algorithms show that the two algorithms are powerful and can obtain good results when applied to segment medical images.

Finding the 3D-Structure of a molecule using genetic algorithm and tabu search methods

The search for the global minimum of a potential energy function is very difficult since the number of local minima grows exponentially with the molecule size. The present work proposes the application of genetic algorithm and tabu search methods, which called GAMCP (Genetic Algorithm with Matrix Coding Partitioning) [4], and TSVP (Tabu Search with Variable Partitioning) [5], respectively, for minimizing the molecular potential energy function. Computational results for problems with up to 200 degrees of freedom are presented and favorable compared with other three existing methods from the literature. Numerical results show that the proposed two methods are promising and produce high quality solutions with low computational costs.

A New Hybrid Particle Swarm Optimization with Variable Neighborhood Search for Solving Unconstrained Global Optimization Problems

Over the past few decades, metaheuristics have been emerged to combine basic heuristic techniques in higher level frameworks to explore a search space in an efficient and an effective way. Particle swarm optimization (PSO) is one of the most important method in meta- heuristics methods, which is used for solving unconstrained global optimization prblems. In this paper, a new hybrid PSO algorithm is combined with variable neighborhood search (VNS) algorithm in order to search for the global optimal solutions for unconstrained global optimization problems. The proposed algorithm is called a hybrid particle swarm optimization with a variable neighborhood search algorithm (HPSOVNS). HPSOVNS aims to combine the PSO algorithm with its capability of making wide exploration and deep exploitation and the VNS algorithm as a local search algorithm to refine the overall best solution found so far in each iteration. In order to evaluate the performance of HPSOVNS, we compare its performance on nine different kinds of test benchmark functions with four particle swarm optimization based algorithms with different varieties. The results show that HPSOVNS algorithm achieves better performance and faster than the other algorithms.

A Survey of Metaheuristics Methods for Bioinformatics Applications

Over the past few decades, metaheuristics methods have been applied to a large variety of bioinformatic applications. There is a growing interest in applying metaheuristics methods in the analysis of gene sequence and microarray data. Therefore, this review is intend to give a survey of some of the metaheuristics methods to analysis biological data such as gene sequence analysis, molecular 3D structure prediction, microarray analysis and multiple sequence alignment. The survey is accompanied by the presentation of the main algorithms belonging to three single solution based metaheuristics and three population based methods. These are followed by different applications along with their merits for addressing some of the mentioned tasks.

Artificial bee colony optimizer for historical Arabic manuscript images binarization

Historical manuscript image binarization is a very important step towards full word spotting system. In this paper, we present a novel binarization algorithm based on artificial bee colony optimizer. The proposed approach contains two phases. The first phase is stretching the intensity level of the image by contrast stretching filter and removing the noise by image cleaning algorithm, the second phase is determining the number of clusters, number of colony and iterations for starting Artificial Bee Colony (ABC) algorithm. The proposed approach is tested on a set of images collected from the electronic Arabic manuscripts database and compared against three famous binarization methods such as Niblacks, Otsus and Savouls. The Experimental results show that the proposed approach is a promising approach and can obtain the desired results better than the other compared methods.

Memetic Artificial Bee Colony for Integer Programming

Due to the simplicity of the Artificial Bee Colony (ABC) algorithm, it has been applied to solve a large number of problems. ABC is a stochastic algorithm and it generates trial solutions with random moves, however it suffers from slow convergence. In order to accelerate the convergence of the ABC algorithm, we proposed a new hybrid algorithm, which is called Memetic Artificial Bee Colony for Integer Programming (MABCIP). The proposed algorithm is a hybrid algorithm between the ABC algorithm and a Random Walk with Direction Exploitation (RWDE) as a local search method. MABCIP is tested on 7 benchmark functions and compared with 4 particle swarm optimization algorithms. The numerical results demonstrate that MABCIP is an efficient and robust algorithm.

A Novel Hybrid Perceptron Neural Network Algorithm for Classifying Breast MRI Tumors

Breast cancer today is the leading cause of death amongst cancer patients inflicting women around the world. Breast cancer is the most common cancer in women worldwide. It is also the principle cause of death from cancer among women globally. Early detection of this disease can greatly enhance the chances of long-term survival of breast cancer victims. Classification of cancer data helps widely in detection of the disease and it can be achieved using many techniques such as Perceptron which is an Artificial Neural Network (ANN) classification technique. In this paper, we proposed a new hybrid algorithm by combining the perceptron algorithm and the feature extraction algorithm after applying the Scale Invariant Feature Transform (SIFT) algorithm in order to classify magnetic resonance imaging (MRI) breast cancer images. The proposed algorithm is called breast MRI cancer classifier (BMRICC) and it has been tested tested on 281 MRI breast images (138 abnormal and 143 normal). The numerical results of the general performance of the BMRICC algorithm and the comparasion results between it and other 5 benchmark classifiers show that, the BMRICC algorithm is a promising algorithm and its performance is better than the other algorithms.