An Agent-Based Approach to the Simulation of Brown Plant Hopper (BPH) Invasions in the Mekong Delta

In this paper we propose an agent-based approach to the modeling and simulation of invasions of brown plant hoppers in the Mekong Delta. Although these seasonal invasions are now heavily studied and their factors quite well known, it is still difficult to predict in advance where and when the next occurrence will appear. As a matter of fact, the conjunction of particular local weather conditions, global winds, biological characteristics of rice crops, agricultural decisions in Vietnam and in the surrounding countries (mainly Cambodia and Thailand), prevent analytical models from delivering interesting results. One of the only options is then to build predictive simulation models, with the constraint that they have to rely on accurate spatial and weather data, and have to represent the knowledge about the various heterogeneous "actors" that play a role in these phenomena (from the insects themselves to the farmers or local authorities). After a short review on similar researches, we argue that agent-based modeling is likely the most interesting approach to fulfill these requirements and we present the methodology we have used to design our model (BPHSim). This model and its implementation are detailed, and followed by an overview of its validation against real data. Then, we discuss the different research directions that this approach will allow us to take in order to tackle the more complicated problem of decision-support when an invasion is predicted.