Multiple paths test data generation based on particle swarm optimisation

Abstract

Considering path coverage as the test adequacy criterion, a modified multiple paths test data generator based on particle swarm optimisation (MMPPSO) algorithm is proposed. During the particle swarm optimisation process, each particle tracks the individual best position and the global best position. For the multiple paths coverage problem, different fitness functions are applied to assess the individual best position and the global best position in MMPPSO. The weighted summation of those branch distance functions is designed as the single path fitness function. The fitness function for the individual best position is the minimum of those single path fitness functions, which guides particles converge to a specific path. The fitness function for the global best position is the summation of those single path fitness functions, which guides the population achieve multiple paths coverage and avoid the premature convergence. The experiments implemented on some benchmarks show that the authors’ approach is more effective and more efficient than other methods, especially for complicated programs and large target path sets.