A genetic programming based iterated local search for software project scheduling

Project Scheduling Problem (PSP) plays a crucial role in large-scale software development, directly affecting the productivity of the team and on-time delivery of software projects. PSP concerns with the decision of who does what and when during the software project lifetime. PSP is a combinatorial optimisation problem and inherently NP-hard, indicating that approximation algorithms are highly advisable for real-world instances which are often large in size. In this work, we propose an iterated local search (ILS) algorithm for PSP. ILS is a simple, yet effective for combinatorial optimisation problems. However, its performance highly depends on its perturbation operator which is to guide the search to new starting points. Hereby, we propose a Genetic Programming (GP) approach to evolve perturbation operators based on a range of low-level operators and rules. The evolution process will go along with the iterated search process and supply better operators continuously. The GP based ILS algorithm is tested using a set of well known PSP benchmark instances and compared with state-of-the-art algorithms. The experimental results demonstrated the effectiveness of GP generated perturbation operators as they can outperform existing leading methods.