Industrial Software Developments Effort Estimation Model
Shahab Nadir, Detlef Streitferdt, Christina Burggraf · 2016
Over the past decade, software has spread to most areas of our lives. This has led to increased demands on product quality and complexity. Industrial software which belong to a safety-critical area where high quality products are essential. Many processes and standards must be completed and met within stipulated deadlines. The complexity of the software and the boundary conditions of developing it dictate the product defect ratio. Requirements will be added and adjusted during the development lifecycle which will leads to changes in the software. All these changes need to be refined and tested to ensure product quality. Predicting software defects and changes is a major part of software engineering. The software defect count is used to assess quality and estimate the effort needed for software fixes, however, it is rarely accurate. Industrial development involves many standards and norms. This research models correction efforts from multiple perspectives. The effort needed for software fixes needs to be combined with overall product effort to ensure software quality. The most common approach is to update the software. Several software update will be applied with each release. Each defect has specific characteristics. These depend on the process and project parameters, with the parameters determining whether a defect exists and the probability of generating a new one. The software process model contained in this thesis will describe these parameters and illustrate their effect on the software defect rate. The software is developed in different phases during the development lifecycle. The effort needed to fix defects depends on the phases where they originated and were detected. This is modeled in this thesis as a defect cost flow. Describing the distribution of defects during these phases helps control quality assurance and limit the effort needed to increase product quality.