Search-based improvement of subsystem decompositions

Abstract

The subsystem decomposition of a software system degrades gradually during its lifetime and therefore it gets harder and harder to maintain. As a result this decomposition needs to be reconditioned from time to time. The problem is to determine a suitable subsystem decomposition that can be used as a basis for future maintenance tasks. This paper describes a new methodology that computes such a subsystem decomposition by optimizing metrics and heuristics of good subsystem design. The main idea is to treat this task as a search problem and to solve it using a genetic algorithm.

Synthesizing Architecture from Requirements: A Genetic Approach

“…This approach only uses one quality factor (reusability), and also a more refined starting point than what is used in our approach. Seng et al (2005) describe a methodology that computes a subsystem decomposition that can be used as a basis for maintenance tasks by optimizing metrics and heuristics of good subsystem design. GA is used for automatic decomposition.…”

Genetic Synthesis of Software Architecture

“…Some approaches also use the tournament technique to select the next generation [Blickle, 1996;Seng et al, 2005].…”

“…Fourthly, the classes that do not exist in the new sequence (and were parts of the deleted subsystem), are collected. Fifthly, the collected classes are distributed to other genes so that all classes will still stay present in the solution [Seng et al, 2005]. The fitness function is formed from a combination of metrics for cohesion, coupling, complexity, cycles and bottlenecks.…”

“…Bottlenecks are subsystems that know about and are known by too many subsystems. A tournament selection is used for selecting the new generation [Seng et al, 2005]. Seng et al [2005] also believe in the building block theorem, and construct their initial population accordingly.…”

“…A tournament selection is used for selecting the new generation [Seng et al, 2005]. Seng et al [2005] also believe in the building block theorem, and construct their initial population accordingly. They bring solutions with high fitness values into the initial population in order to ensure the presence of good building blocks from the very beginning.…”