Applications of fuzzy logic to software model / Yuhanim Hani Yahaya

Yuhanim Hani, Yahaya (2000) Applications of fuzzy logic to software model / Yuhanim Hani Yahaya. Undergraduates thesis, University of Malaya.

Preview

PDF (Academic Exercise (Bachelor’s Degree) Download (35Mb) | Preview

Abstract

This thesis investigates the use of fuzzy logic approach in software metrics application. Software metrics defines a standard way of measuring the properties of software products, development processes and resources. The most common application of software metrics is the software cost estimation where it predicts the effort required for completing certain stages of a software development life cycle. The ability to obtain an accurate effort prediction is essential for the cost estimation process, as it helps a project manager to specify the efforts needed for project development. In relation to this, cost estimation models such as COCOMO (Constructive Cost Model), Function Points and SLIM (Software Life-Cycle In Management) which used specified equations for estimating development effort have been proposed. However, these models suffer several limitations in terms of the inputs. Inputs to these models may include experience of the programmer, the required reliability of the software, complexity of the project and an estimate of the project size. These inputs are subjective (non-numerical) and thus requiring expert knowledge. Some of these inputs are confusing and are not known with reasonable degree of certainty until the project is completed. In an attempt to overcome this problem and to fulfil the needs, fuzzy logic has been studied for effort prediction. Fuzzy logic is a form of logic that deals with subjective and uncertainty values. It allows expert knowledge to determine the input values by using the linguistic terms instead of mathematical equations. Project managers are in fact able to classify the inputs by using linguistic term such as "High level of project complexity" and "Size of the project is medium". The Fuzzy Logic Effort Prediction (FLEP) program is produced in an attempt to mode the fuzzy logic approach in estimating effort development. Using Mean Magnitude of Relative Error (MMRE) measurement, a comparison between fuzzy logic and other techniques have been made. Four cost estimation techniques are compared, COCOMO, Function Points, SLIM and Fuzzy Logic. The results showed that fuzzy logic technique provides better estimation for development effort in the early stage of development life cycle.

Actions (For repository staff only : Login required)