Software Performance Estimate using Fuzzy Based Backpropagation Learning

Introduction

With the rapid development of computer technology, it is becoming increasingly important in everyday activity of an individual resulting in increase of dependability on computer technology. Therefore, every individual uses a software application based on the different parameters with performance being of high priority. An individual starts relying on software based on the performance which makes it very important from user’s perspective. Software performance prediction modeling techniques are the basis of quantitative analysis of software and have caught the focus of software developers. It has becomes more difficult to conduct quantitative and qualitative performance prediction and analysis for software systems because of the increasing scale and complex structures and behaviors. However, fuzzy and backpropagation learning techniques are of great help in dealing with this issue (Takacs, 2004) (Sram & Takacs, 2011) (Mitra & Hayashi, 2000). The determination of software performance indices is the crucial task for software engineers across the globe. The quality indices are not only the basis for the software performance design and the constraints during the software development cycle (Piattini, M., M., Jiménez & L, 2001). Software performance is usually separated into quantitative and qualitative index. Quantitative index is quantified software performance parameter values, such as reliability of software are defined as the likelihood of breakdown free operation of a software program for a specific time based on certain conditions (Hsu & Huang, 2011). Most of the research in software components is devoted to specification, design, reuse, and cataloguing of the components. Software reliability and performance is defined as the probability of the failure free operation of a software system for a specified period of time in a specified environment (Bontempi & Kruijtzer, 2002) (“Software engineering fundamentals,” 1996). With the increase in complexity of software applications more emphasis is given on reusability and function library management (Jakubek & Strasser, 2004).

Software application performance is defined as follows:

• •

  The likelihood of a given system performing its task adequately for a particular period of instance under the accepted operating conditions.

• •

  The likelihood that software will provide breakdown-free function in a preset environment for a predetermined interval of time.

Software performance is normally deliberate per unit of time, whereas likelihood of breakdown is generally time independent. Software performance differs considerably from program correctness which paves way to independent effort estimation method for software performance. Fuzzy theory provides the desired result in estimating the software attributes regarding its technical or non-technical attributes (Shi & Mizumoto, 2000) (Yang, Xu & Zang, 2000) (P, Liu, Hongxing & L, 2004) (Kuo, J, Wu, P, Wang & P, 2002).