Optimization of Automated Software Contracts Generation by Modified Particle Swarm Optimization

Abstract

Incorrect software behaviours are detected using a combination of Programming language contracts
and alliedresources of programs namely static source code Analyzers and Runtime verifyingframeworks.
To surpass therestrictions of inordinate run-time overheads in the course of verification of software
modules with timing constraints, Optimization of the contracts after their automated development is
imperative. This is carried out by retrieving the structural and behavioural dependency information of the
real time multithreaded Java source code under test, then remodelthem as contract details in a Decision
tree after Static and Dynamic analyses. The next step incorporates transformation of the conditions into
contracts. The Optimizer logic is implemented as a modification of Particle Swarm Optimization
technique during this transformation by adopting an objective fitness function that substantiallylends to
handle only the source files that contain valid decisions. Evaluation measures specifically processing
time for contracts formation, Memory and CPU Utilization, Number of files processed are used to
determine the effectiveness of the proposed system comparable with no optimization applied.