## SIMULATION ON EVOLUTION OF COMMUNICATION CONVENTIONS

Your browser does not support Java, so nothing is displayed.

### Explanation

In this simulation we consider a scenario of 400 communicative individuals in an unbounded, torus shaped space. This scenario is similar to the one described in Oliphant (1994), the main difference is that each agent goes through 5 life stages; after the fifth the first follows. During the fist stage, an individual selects a teacher from the set of its neighbours according to the communication success (fitness) they have accumulated during the previous life stage (a neighbour with double fitness has double probability to be selected). The newborn agent then acquires with a certain precision the communication system of its teacher: each mapping in the teacher's communication system is learned with probability 1-M and it is randomly chosen with probability M (where M is the language acquisition inprecision or mutation rate).

The left side panel represents the population of agents. The color of each square characterizes the communication system of the associated agent. The bottom left number is the age and the top right number is the fitness rounded and multiplied by 10 (X stands for 10). The graph in the right side panel represents the frequency of each colored communication system for each iteration. The black line specifies the level of the average fitness of the entire population.

The right side of the panel containes different parameters which can be set accordingly:

• Iterations: number of life cycles the simulation runs.
• Meanings: number of meanings used by the agents, equivalent to the number of forms (symbols) they are able to express.
• Mutation: the mutation rate characterizing the acquisition phase.
• Random Positioning: whether the agents are randomly displaced at each iteration (ON) or stay in the same position during the life cycle (OFF).
• Optimal Filling: whether in the starting point of the simulation each agent is initialized with a unique optimal communication system (ON) or with a random one (OFF).
• Optimal Communication: all the meanings are used in the communication between two agents. Furthermore the payoff of the communication is maximum if and only if all the meanings are correctly communicated (0 otherwise).
• Iterative: when active each agent has two possible communication systems: in each communication attempt each agents uses the first communication system if the previous reception was successful, the second if unsuccessful. Each communication between two agents is iterated 16 times.
• Payoff: when communicating to its neighbours, each agent cumulates a transmission and reception success rate. The payoff matrix is specifies by 4 numbers a,b,c,d: -1 &le a,b,c,d &le 1 specifying how much to reward each agent for
1. successfull transmission
2. unsuccessfull transmission
3. successfull reception
4. unsuccessfull reception
For more details you can download the technical report of this project.

### References

Oliphant, M. (1994). The dilemma of Saussurean communication. BioSystems, 37(1-2), 31-38.