Convergence analysis and performance of an improved gravitational search algorithm

- IGSA that employs personal best memory and chaotic perturbation is proposed.
- Stability condition of the IGSA is analyzed based on linear system theory.
- Global convergence of the IGSA is presented using stochastic process method.
- Sensitivity analysis of parameters for IGSA is discussed on test functions.
- Comparative experiments are performed to evaluate performance of the IGSA.
- Results show IGSA outperforms other heuristic algorithms and GSA variants.