An Agent-based Model of the Anopheles gambiae Distinct Life Stages and its C++ implementation: AGILESim

Ying Zhou(1), S. M. Niaz Arifin(1), James Gentile(1), Steven J. Kurtz(1), Gregory J. Davis(1), Barbara A. Wendelberger(2), Greg Madey(1)

(1) Department of Computer Science and Engineering and (2) Department of Biological Sciences University of Notre Dame, Notre Dame, IN 46545

TITLE:An Agent-based Model of the Anopheles gambiae Distinct Life Stages and its C++ implementation: AGiLESim

ABSTRACT: Malaria is a mosquito-borne infectious disease. Each year, it kills around two million people, and most of them are young children in Sub-Saharan Africa. Anopheles gambiae transmits the most dangerous malaria parasite of Plamodium falciparum among humans. Due to its critical role in the malaria transmission, modeling the Anopheles gambiae population dynamics and behavior can help understand why, when, and how malaria can become prevalent in an area, and assist in deploying cost-effective malaria vector control strategies.

This abstract introduces an agent-based model of the Anopheles gambiae mosquito life cycle and its C++ implementation called AGiLESim. Our model tracks the Anopheles gambiae's eight distinct life stages: egg (e), larva (l), pupa (p), ImmatureAdult(IA), MateSeeking (MS), BloodmealSeeking (BMS), BloodmealDigesting (BMD), and Gravid (egg-carrying) (G), and the transitions among these stages.

In our model, there are two kinds of agents: mosquito and habitat. Each mosquito agent is distinguished by a number of fundamental characteristics: age, sex, genotype, stage, and so on. It also has some specific behaviors such as mating, stage-updating, and ovipositing. Each habitat agent is characterized by attributes, including ecological factors such as carrying capacity, the egg population, the larval population, and so on. A habitat agent also employs the adding, killing, and updating functions to maintain the aquatic mosquitoes living in it.

Three major interactions determine how agents in our model communicate with or affect other agents: 1) the mating interaction between a female and a male both at MS stage; 2) the second interaction is between an ovipositing female adult and the aquatic residents in a habitat. It affects how many eggs a female prefers to lay in this habitat and its subsequent ovipositing behavior. Such a mosquito would become less "picky" and prefer to lay the remaining eggs into future habitats; 3) larvae in the different age groups in one habitat also interact with each other. Mosquito agents' behaviors are guided by three major rules: the daily mortality rate (DMR) and mosquito development rules for all mosquitoes at the eight stages, and the oviposition rule for gravid female agents. Habitat agents’ rules are responsible for maintaining their state variables and the aquatic mosquito agents within them.

The C++ implementation, AGiLESim, consists of six modules. The time step used in AGiLESim is one day. Module 1 performs the simulation initialization and is called only once at the beginning of a simulation. At each time step, five modules are called sequentially. Module 2 is in charge of stochastically selecting some adult mosquitoes from each adult age group to kill and then removing them from the system based on the adult daily mortality rate rule of this group. Module 3 deals with stochastically picking some eggs, larvae from each larval age group, and pupae in each habitat according to their respectively daily mortality rate rules. Module 4 updates the aquatic agents that survived module 3. Module 5 updates the adult agents that survived module 2. Lastly, module 6 collects the data. AGiLESim takes an arbitrary number of adult agents at IA stage as one input that consists of half females and half males. Other inputs include arbitrary numbers of habitat agents with adjustable carrying capacities, a set of parameter values used in the rules, and the weather profile. AGiLESim outputs the simulation data as follows: 1) the male mosquito abundance and the female mosquito abundance; 2) the egg, larval, and pupa populations in each habitat; 3) the number of female mosquitoes older than 12 days; 4) the female adult population age structure.

We will continue to refine our model and make it more representative of the mechanisms underlying Anopheles gambiae population dynamics. We will also introduce widely used malaria control interventions such as Insecticide-treated nets (ITNs) into our model and investigate prevalent mosquito insecticide resistance. Incorporating agent spatial properties such as mosquito agent movement and habitat agent coordinates is also one of the next steps in our development plan.

Back to Swarmfest2010_program