ITGO: Invasive tumor growth optimization algorithm

• A new heuristic optimization algorithm inspired by tumor growth was proposed.
• A new mechanism and different search strategies was proposed.
• Proposed algorithm was compared with the well-known algorithms such as ABC, DE, PSO.

This paper proposes a new optimization algorithm named ITGO (Invasive Tumor Growth Optimization) algorithm based on the principle of invasive tumor growth. The study of tumor growth mechanism shows that each cell of tumor strives for the nutrient in their microenvironment to grow and proliferate. In ITGO algorithm, tumor cells were divided into three categories: proliferative cells, quiescent cells and dying cells. The cell movement relies on the chemotaxis, random walk of motion and interaction with other cells in different categories. Invasive behaviors of proliferative cells and quiescent cells are simulated by levy flight and dying cells are simulated through interaction with proliferative cells and quiescent cells. In order to test the effectiveness of ITGO algorithm, 50 functions from CEC2005, CEC2008, CEC2010 and a support vector machine (SVM) parameter optimization problem were used to compare ITGO with other well-known heuristic optimization methods. Statistical analysis using Friedman test and Wilcoxon signed-rank statistical test with Bonferroni–Holm correction demonstrates that the ITGO algorithm is better in solving global optimization problems in comparison to the other meta-heuristic algorithms.

A new swarm intelligence algorithm for continuous optimization problem is performed according to the mechanism of a real world tumor growth. Experiment shows the search trajectory of the three different types of tumor cells (Pcell, Qcell, Dcell) for a shifted unimodal function with noise.Figure optionsDownload as PowerPoint slide