S. Masrom

Time-Varying mutation in particle swarm optimization

One of significant improvement for particle swarm optimization (PSO) is through the implementation of metaheuristics hybridization that combines different metaheuristics paradigms. By using metaheuristics hybridization, the weaknesses of one algorithm can be compensated by the strengths of other algorithms. Therefore, researchers have given a lot of interest in hybridizing PSO with mutation concept from genetic algorithm (GA). The reason for incorporating mutation into PSO is to resolve premature convergence problem due to some kind of stagnation by PSO particles. Although PSO is capable to produce fast results, particles stagnation has led the algorithm to suffer from low-optimization precision. Thus, this paper introduces time-varying mutation techniques for resolving the PSO problem. The different time-varying techniques have been tested on some benchmark functions. Results from the empirical experiments have shown that most of the time-varying mutation techniques have significantly improved PSO performances not just to the results accuracy but also to the convergence time.

Hybridization of Particle Swarm Optimization with adaptive genetic algorithm operators

Particle Swarm Optimization (PSO) is a popular algorithm used extensively in continuous optimization. One of its well-known drawbacks is its propensity for premature convergence. Many techniques have been proposed for alleviating this problem. One of the alternative approaches is hybridization. Genetic Algorithms (GA) are one of the possible techniques used for hybridization. Most often, a mutation scheme is added to the PSO, but some applications of crossover have been added more recently. Some of these schemes use adaptive parameterization when applying the GA operators. In this work, adaptively parameterized mutation and crossover operators are combined with a PSO implementation individually and in combination to test the effectiveness of these additions. The results indicate that an adaptive approach with position factor is more effective for the proposed PSO hybrids. Compared to single PSO with adaptive inertia weight, all the PSO hybrids with adaptive probability have shown satisfactory performance in generating near-optimal solutions for all tested functions.

Backpropagation neural network with new improved error function and activation function for classification problem

Neural network has been used extensively for classification and many real world applications. The most commonly used neural network is multilayer perceptron with backpropagation (BP) algorithm. However the major problem of this algorithm is slow convergence rate and trap to local minima. The convergence is dependent on network parameters such as learning rate, momentum term and slope of activation function as well as its error function. This study proposes a New Improved BP algorithm which applies adaptive activation function using arctangent function in input-to-hidden layer and sigmoid logistic function in hidden-to-output layer. The efficiency and accuracy of the new improved method have been implemented and tested on two benchmark datasets: XOR and Balloon. The results show that the proposed method improved the convergence speed. However the classification accuracy is not very encouraging.

Dynamic parameterization of the particle swarm optimization and genetic algorithm hybrids for vehicle routing problem with time window

Particle Swarm Optimization PSO is a well known technique for solving various kinds of combinatorial optimization problems including scheduling, resource allocation and vehicle routing. However, basic PSO suffers from premature convergence problem. Many techniques have been proposed for alleviating this problem. One of the alternative approaches is hybridization. Genetic Algorithms GAs are one of the possible techniques used for hybridization. Most often, a mutation scheme is added to the PSO, but some applications of crossover have been added more recently. Some of these schemes use dynamic parameterization when applying the GA operators. In this work, dynamic parameterized mutation and crossover operators are combined with a PSO implementation individually and in combination to test the effectiveness of these additions. The results indicate that all the PSO hybrids with dynamic probability have shown satisfactory performance in finding the best distance of the Vehicle Routing Problem With Time Windows.

A taxonomy of low-level hybridization in metaheuristics algorithms

In the last two decades, a lot of metaheuristics approaches have been discovered to tackle large-scale of combinatorial optimization problems. Among the approaches, one of the most effective is so-called metaheuristics hybridization that tries to combine different strengths of different algorithms. In hybridization techniques, implementing low-level hybridization is considered as the most complicated due to the internal structure modification of the hybrid algorithms. In addition, different components of the hybrid algorithms are strongly inter-dependent and they must fit will together in solving a particular problem. Therefore, determining appropriate components to be retained and dropped or replaced in each of metaheuristic algorithm is a very difficult task. Responding to the complexity, this paper presents a new taxonomy for low-level hybridization. Then, a review of several implementations for low-level hybridization in metaheuristics is given with regards to the taxonomy. The outcome of study is useful in providing guidance for effective implementation of low-level hybridization.

Low-level Teamwork Hybridization for P-metaheuristics: A review and comparison

Inspired by nature, many types of Population based metaheuristics or P-metaheuristics is cropping out of research labs to help solve real life problems. Since every metaheuristics has its own strength and weaknesses, hybridizing the algorithms can sometimes produce better results. To this date of literature, Low-level Teamwork Hybridization is considered as an effective and popular method for hybridization of P-metaheuristics. In many cases however, the approach might prove to be quite complicated. The hybridization often requires metaheuristics internal structure modification in order for the different algorithms to fit well together. Another difficulty is in determining which strategies to be retained and which to be dropped or replaced in each of the metaheuristic algorithms. This paper provides a general abstraction for P-metaheuristics and describes the main P-metaheuristics components that are suitable candidates for hybridization. The review and comparative study of several implementations of Low-level Teamwork Hybridization is also presented.

Non-repudiation in order, delivery and payment process for a sustainable online business

Online businesses in Malaysia suffer from the lack of non-repudiation in the ordering, delivery and payment process. Most businesses cannot sustain their growth because of this limitation. In this paper we propose a new model for the implementation of online business. The solutions can also be adapted to any country for the most part except for the delivery process that requires the use of a smartcard-based national identity card. The non-repudiation evidences that are generated from the system can be used by both the consumer and the merchant to resolve possible future disputes. The evidence is also admissible in the court of law since they are generated from the Public Key Infrastructure (PKI) that is recognized by the authority as legally binding.

Scripting language constructs for dynamic parameterizations of PSO-GA hybrids

Responding to the difficulties of implementing meta-heuristics hybridization, this paper introduces a set of scripting language constructs for the rapid algorithm design and development focusing on the two well-known metaheuristics, namely Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). Additionally, the PSO-GA hybrids have been embedded with dynamic parameterization. In this paper, the compiler specification and codes for developing the scripting language constructs are described. Then, based on the several algorithms of PSO-GA hybrids that have been developed with the scripting language constructs, the Line of Codes (LOC) are measured in order to test the easiness of the programming language. The results show that across all algorithms, the scripting language is anticipated to enable easy programming, which has been presented by the very less of LOC compared to the JAVA programming language.

Easy and Concise Programming for Low-Level Hybridization of PSO-GA

Responding to the difficulties of implementing Low-Level Hybridization (LLH) of meta-heuristics, this paper introduces a reusable software for the algorithm design and development. This paper proposes three implementation frameworks for the LLH of Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). Then, with attempt to support a more effective programming environment, a set of scripting language constructs based on the proposed implementation frameworks is developed. For evaluation, twelve algorithms that composed of nine LLHs and three single PSO have been coded and executed with the scripting language. The results demonstrate that the scripting language is anticipated for enabling of an easier and more concise programming for effective rapid prototyping and testing of the algorithms.

Software Framework for Flexible User Defined Metaheuristic Hybridization

Metaheuristic algorithms have been widely used for solving Combinatorial Optimization Problem (COP) since the last decade. The algorithms can produce amazing results in solving complex real life problems such as scheduling, time tabling, routing and tasks allocation. We believe that many researchers will find COP methods useful to solve problems in many different domains. However, there are some technical hurdles such as the steep learning curve, the abundance and complexity of the algorithms, programming skill requirement and the lack of user friendly platform to be used for algorithm development. As new algorithms are being developed, there are also those that come in the form of hybridization of multiple existing algorithms. We reckon that there is also a need for an easy, flexible and effective development platform for user defined metaheuristic hybridization. In this article, a comparative study has been performed on several metaheuristics software frameworks. The result shows that available software frameworks are not adequately designed to enable users to easily develop hybridization algorithms. At the end of the article, we propose a framework design that will help bridge the gap. We foresee the potential of scripting language as an important element that will help improve existing software framework with regards to the ease of use, rapid algorithm design and development. Thus, our efforts are now directed towards the study and development of a new scripting language suitable for enhancing the capabilities of existing metaheuristic software framework.