TDK conference 2017

Many animal groups such as schools of fishes or flocks of birds clearly display structural order often resulting in complex patterns. A fundamental question that what is the adaptive value of these patterns from the perspective of the individuals in a group. Our objective was to assess whether the alignment of the individual motion to other group members and the resulting collective level patterns can be adaptive (can maximize the individual fitness) in such groups where the members must balance two competing needs: (1) the maximization of individual foraging rate and (2) maintain a minimum distance between individuals at all times to avoid collisions.

In our model, the individual motion was determined by a so-called strategy parameter expressing the relative influence of collision-avoiding and food-looking motivation. By measuring the actual collision and food intake rates we determined a fitness function which was weighted by parameter w expressing the relative effect of the collision and food intake on the fitness function. We explored the collective patterns produced this system with the help of computer simulations at different densities.

We found that between a critical interval of cost of collision (0.30.75) . The maximum fitness resulting strategies at different densities (d) were: d=0.4: s=0.95, d=2: s=0.9, d=2.8: s=0.9. This was the result of the higher collision rates compared with food intake rates at strategy parameter levels which did not generated flocking patterns. While in highly synchronized moving groups the collision rates were practically zero that counterbalanced the reduced food intake rates.

To sum up, in such groups where the fitness was determined by both the food intake efficiency and the costs of collisions following a flocking behavior that resulted in highly aligned patterns was an optimal strategy by which the group members could minimize the fitness costs of collisions that could correspond to an evolutionary optimum in natural biological systems too. We intend to use this model as a reference for our perspective research where the individual rules of the motion are not fixed but determined by an adaptable neural network.