Balancing human livability and bumblebee population in urban green areas via Agent-Based Model and Evolutionary Algorithms

We present an agent-based model (ABM) for simulating the ecological dynamics within urban parks, specifically focusing on bumblebee demography and conservation. The objective of the model is to simulate the impact of park management on the population of new bumblebee queens. The ABM features simulated agents for plants, bumblebees, and bumblebee colonies, incorporating insights from a literature review on bumblebee behavior and plant ecology. It operates over a three-year simulation period, during which the effects of four key parameters are explored, namely the percentage and the shape of unmowed areas, the mowing frequency, and pesticide treatments. The significance of the ABM becomes evident in estimating the impact of various park management strategies on the bumblebee population. By employing NSGA-II, we showcase a practical application of the proposed ABM, generating a Pareto front that delineates the trade-offs between park livability and the anticipated number of new queens. We also investigate how different solution representations yield different Pareto fronts and how different combinations of initialization methods, population size, and number of generations influence the distribution of the final population. Overall, the proposed approach can offer valuable insights to urban environment managers, aiding in the delicate balance between park livability and biodiversity conservation.