Optimization in Software Engineering -- A Pragmatic Approach

TL;DR

This paper provides a pragmatic overview of how optimization techniques are applied in software engineering, emphasizing practical usage, decision support, and validation across the software development lifecycle.

Contribution

It offers a comprehensive overview of pragmatic optimization practices in software engineering, including a process checklist and a case study example.

Findings

Optimization supports decision-making in software engineering.

A checklist facilitates effective optimization application.

Return-on-Investment analysis guides problem scope and effort.

Abstract

Empirical software engineering is concerned with the design and analysis of empirical studies that include software products, processes, and resources. Optimization is a form of data analytics in support of human decision-making. Optimization methods are aimed to find the best decision alternatives. Empirical studies serve both as a model and as data input for optimization. In addition, the complexity of the models used for optimization trigger further studies on explaining and validating the results in real-world scenarios. The goal of this chapter is to give an overview of the as-is and of the to-be usage of optimization in software engineering. The emphasis is on pragmatic use of optimization, and not so much on describing the most recent algorithmic innovations and tool developments. The usage of optimization covers a wide range of questions from different types of software engineering problems along the whole life-cycle. To facilitate its more comprehensive and more effective usage, a checklist for a guided process is described. The chapter uses a running example Asymmetric Release Planning to illustrate the whole process. A Return-on-Investment analysis is proposed as part of the problem scoping. This helps to decide on the depth and breadth of analysis in relation to the effort needed to run the analysis and the projected value of the solution.