An Evolutionary Constraint-Handling Technique for Parametric Optimization of a Cancer Immunotherapy Model

Recent studies have shown that evolutionary constraint-handling techniques are capable of solving optimization problems with constraints. However, these techniques are often evaluated based on benchmark test functions instead of real-world problems. This paper presents an application of evolutionary constrained parametric optimization for a breast cancer immunotherapy model formulated based on biological principles and limited clinical results. It proposes a new constraint-handling technique that partitions the population into different sections to enhance the evolutionary search diversity. In addition, the upper bound of each section is reduced dynamically to drive the convergence of individuals toward the feasible solution region. Experimental results show the effectiveness and robustness of the proposed constraint-handling approach in solving parametric optimization problems. Moreover, the evolutionary optimized cancer immunotherapy model can be used for prognostic outcomes in clinical trials and the predictability is considered significant for such a parametric optimization approach.