By Ioulia Fenton and Tracey Li*
By now you have probably already noticed that SimPachamama is based on something called Agent-Based Modeling (ABM). And you may very well be scratching your head in confusion, so let us explain.
ABM is the term given to computer models which simulate the behavior of individual ‘agents’. An agent is any entity which can be considered to behave autonomously, assessing its own individual situation and making decisions based on this situation according to a set of rules. If we want to explore the behavior of a single community, then each agent could be an individual person within that community. But it would also work if an agent is a group of people – say an entire village – if we wanted to study the changes in a large geographical region containing several interconnected villages. Deciding whether to treat a single person, a whole community, or even some larger group, as an agent depends on what you want to study.
The agents make decisions based on certain ‘rational’ rules. Let’s think about a poor farmer in Bolivia who has been given 50 hectares of forest in the Amazon for him or her to start a new life. This is in fact what happens in Bolivia through government programs, which relocate families from areas that are crowded to less populated parts of the country. Whether that farmer clears the land to plant crops or keeps the forest intact is assumed to depend on what is more profitable. Thus, the rule governing his or her decision is:
¨If I can make more money this year from clearing forest to grow and sell corn than the money I would receive from the government to conserve the forest, then I will clear the forest.¨
In a game like SimPachamama which tries to simulate the real world as closely as possible, this kind of rule is generated from real data and then added to the coding of the game. Typically, the simulation will contain many such rules, taking into account different situations that can affect how the agents behave. The simulation is then run to see what would happen to the agents and their environments when lots of rule-based decisions happen over an extended period of time. SimPachamama runs for 20 years at a time and the mayor tweaks different policies to try to affect the economic and social wellbeing of the community, while enticing them to leave more of the forest standing.
Although the behavior of each agent may be relatively simple to model and describe, the behavior of the entire system (i.e. all the agents) is not easily predictable, especially if there are many agents involved with each one taking many possible decisions. One of the strengths of ABM is that it allows you to see something called emergent phenomena. These are trends and effects which arise as a result of the interactions between individual agents, rather than as a result of each individual’s actions. As such, they are a property of the entire system, and not the individual agents, because the phenomena only emerge when there are many agents together; ¨the system is greater than the sum of its parts.¨ This makes emergent phenomena difficult or often impossible to predict, based on knowledge of only the individual agents. As more and more interactions are included between agents, the behavior of the system becomes increasingly complex.
In SimPachamama each agent is an individual farmer (or the farmer plus their family), and the whole system is the community in which they live. Several decisions are built into the game. As mentioned, since the aim of the game is to enact policies that reduce deforestation and increase wellbeing, as in real life, a key decision of each farmer is whether or not to chop down forest and plant crops. If a family is not growing enough to live well, they may decide to earn a little extra money by working on someone else’s farm (sell their labor) where the owner needs an extra hand (buys their labor). If the going gets tough in the community, some people may migrate to the city from the countryside in order to find work. When prospects improve back home, they – or other folk from the city – may migrate the other way. And once a family has saved some money, they may then take the decision to invest in a cow in order to start the more profitable and less labor intensive practice of cattle ranching.
These decisions are affected by the policies that the mayor (you, the player) implements, such as putting into place deforestation taxes, paying community members to conserve forests, creating non-agricultural ‘green jobs’, and accepting a cash injection from other countries to help you reduce deforestation. The policies work together to make either the conservation of forest or the clearing of forest more attractive to different people at different times. The outcomes of these policy options – the emerging phenomena – may not be immediately obvious and it may take you several attempts before you succeed in making your community happy and wealthy and your environment forested and healthy.
The good news is that, while all this may sound official and alien, it is actually quite simple and anyone (including you!) can learn the basics of agent-based modeling and then create his or her own interactive simulations using a variety of free platforms. NetLogo is one of the most user-friendly ones for beginners and includes a library of example models and a comprehensive manual to help you get started. This is also the platform on which SimPachamama is built, and since we have made the coding freely available to everyone—once you’ve grasped the basics of agent-based modeling and NetLogo—you can begin to modify the game to suit your own needs and interests.[contact-form to=’firstname.lastname@example.org’ subject=’SimPachamama: What is ABM?’][contact-field label=’Like this article? Sign up for weekly updates from INESAD’ type=’email’ required=’1’/][contact-field label=’Your name’ type=’name’ required=’1’/][/contact-form]
Ioulia Fenton leads INESAD’s international communications and outreach and Tracey Li is a research and communications associate with INESAD.