An Enhancement of Javadi Et. Al.’s Non-Dominated Sorting Genetic Algorithm-II-Grid-Based Crowding Distance Algorithm (NSGA-II-Gr) For Resource Allocation Applied in Optimizing Rabi Crops Yield

Agriculture serves as the backbone of the Philippine economy, with its ability to sustain a growing global population reliant on effective resource management. To optimize resources such as fertilizers, pesticides, and area, researchers have employed Javadi Et. Al.'s Non-Dominated Sorting Genetic Algorithm-II-Grid-Based Crowding Distance Algorithm (NSGA-II-Gr). Resource allocation and optimization often involve multi-objective decision-making, requiring careful trade-offs among competing parameters. However, upon simulating in higher dimensions, it converges prematurely to suboptimal solutions. After several iterations, the population tends to be dominated by the 'best' solution, leading to premature convergence. This reduces the diversity of candidate solutions and increases the risk of converging to a local optimum, ultimately limiting the exploration of solutions. To address this issue, the researchers introduced a separate spreading mechanism where the mutation intensity decreases over generations. Initially, the mutation strength is high to ensure significant diversity while later in the process, it weakens to promote convergence. This enhancement successfully allowed the algorithm to first explore the solutions before converging. By effectively preventing premature convergence, the modified algorithm gains the ability to explore a broader range of potential solutions. This advancement is particularly valuable for optimizing crop yield and identifying the most effective combinations and trade-offs in agricultural resource management.