Trends and Techniques of Software Cost and Performance Measurement

Abstract

Software measurement is an evolving aspect of software engineering providing the necessary support to plan, execute, control, and improve the software development process. However, Software metrics constitute a set of acceptable variables of measurement encompassing diverse activities that require some level of quantification of which cost, and effort estimation are inclusive. This study was aimed at reviewing the various trends and techniques in software cost and performance measurement. The study provided an overview on existing trends and techniques of software cost and performance measurement. The method adopted in the study is the meta-synthesis systematic literature review which is a non-statistical oriented technique used to analyze, integrate and interpret the outcome of numerous qualitative research studies. For investigation, the techniques’ strengths and flaws were identified and contrasted. Algorithmic and non-algorithmic cost estimation approaches were discovered in the study. The algorithmic technique explores the COCOMO, Putnam and the Function Point Analysis models while the non-algorithmic technique explores the Neural Networks, Particle Swarm Optimization, Genetic Programming, Fuzzy Logic, Expert Judgment, Estimating by Analogy, Top-Down Method, and the Bottom-up Estimating Methods. The comparative results show that there is no single best method as their pros and cons are complementary to one another. A combination of two or more models, on the other hand, is expected to produce better results.