Article Preview
TopIntroduction
Agent-based modeling (ABM) can be applied to a domain with or without an explicit representation of space. In some cases, however, an explicit spatial representation may be required for certain aspects of the ABM to be modeled more realistically. For example, in a spatial ABM of malaria, events like obtaining a successful bloodmeal (host-seeking) or finding an aquatic habitat to lay eggs (oviposition) can be modeled by utilizing the distribution of corresponding resources in the landscape.
Malaria is one of the top three pathogen-specific causes of global mortality, causing an estimated one million deaths per year, mainly in children (WHO, 2011). Only female mosquitoes of the genus Anopheles transmit human malaria, and as such are known as malaria vectors. The species Anopheles gambiae is the most important malaria vector in Sub-Saharan Africa, and one of the most efficient vectors (in terms of malaria transmission) in the world. Earlier, we developed an agent-based model derived from a conceptual entomological model of the A. gambiae lifecycle (Zhou et al., 2010; Arifin et al., 2010a; Gentile, Davis, StLaurent, & Kurtz, 2010). The model, however, was non-spatial: none of the agents and/or environments possessed any spatial attributes.
In this study, we describe a spatial extension of the previous model. Though in both models, all mosquito agents are represented individually, in the new spatial model, the agents also possess explicit spatial information. We show how the previous model and the current spatial model yield consistent results with identical parameter settings (whenever applicable), and hence are docked. We also show how spatial heterogeneity affects some results in the spatial model.
Spatial heterogeneity is considered as one of the most important factors for an effective representation of the environment being modeled. In the discipline of spatial epidemiology (also known as landscape epidemiology), in most cases, the probability of disease transmission significantly declines with distance from an infected host. Thus, the spatial locations of pathogens, hosts and vectors are fundamentally important to disease dynamics (Ostfeld, 2005).
In modeling malaria with ABMs, representation of space may be crucial (Gu, 2009a, 2009b; Menach, 2005). The dynamics of malaria can be subject to substantial local variations that result from various spatial differences (Vries, 2001). Examples of local variations may include locations of aquatic habitats and bloodmeal events, characteristics of mosquitoes, etc. For malaria models, space can be represented as mosquito world, aquatic habitats, etc. for the mosquito agents; and as houses, huts, etc. for the human agents.
In our malaria simulation, some events (e.g., host-seeking, oviposition) by nature require spatial attributes. The underlying spatial heterogeneity defines the spatial distribution of resources, and controls how easily adult female mosquitoes may find resources that are necessary to complete their gonotrophic cycle (the cycle of obtaining bloodmeals and ovipositing eggs). This, in turn, directly affects the mosquito population in the ABM (Arifin, Davis, & Zhou, 2011).