Solving the Components Identification Problem using Ant Based Colony Technique

Abstract

Component-based software engineering (CBSE) is defined as the design and development of newlinesoftware using reusable software components. It reduces the cost of developing software, newlineincreases developers productivity, and reduces the time needed to develop software. In newlinecomponent-based development, the software is partitioned to identify components. Identification newlineof components is a crucial task. Several methods are proposed for this. Most of the proposed newlinemethods use the clustering approach. It is studied that the component identification problem is an newlineoptimization problem. So optimization algorithms can be used to solve this problem. This thesis newlinepresents a novel method of Ant-based Component Identification (ABC_I) for component newlineidentification based on an evolutionary approach. Logical components are considered here. A newlinelogical component is a component representing requirements accepts for technology, newlineenvironment, and constraints. During component-based development, a system is portioned into newlinelogical components. Component identification is done using the use case model. So directly from newlinethe requirement analysis, components can be identified. Cohesion, coupling, complexity, Actors newlineof the use case model, and use cases are used to calculate the fitness value. It automatically newlinecalculates the number of components. ABC-I uses Ant Colony Optimization. The proposed newlinemethod is evaluated using three real-world system use case models. A result shows that ABC-I is newlinesuitable for the identification of components. The performance is better than the existing newlineclustering techniques and genetic algorithm (GA) based techniques. newline