Notes taken by Tobias Langner

Optimal Pheromone Utilization

Roman Kecher presented joint work with Yehuda Afek and Moshe Sulamy in which they consider the decidability of an abstract treasure searching problem. A set of k mobile agents are initially located in the origin of an infinite grid and the goal is to find an adversarially located treasure. The execution of an algorithm terminates when at least one agent is located in the same grid cell as the treasure. All agents execute the same protocol and the protocols and the agents are oblivious to the input parameters k and D, where D is the distance to the treasure.

Roman explained, that -- contrary to the previous talk by Tobias Langner -- the agents can only communicate using pheromones: They are allowed to leave permanent pheromone traces in any grid cell, which any other agent is able to sense for the rest of the execution. No other means of communication are allowed.

As first result, Roman presented a lower bound showing that at least \Omega(k) pheromones traces are needed to locate the treasure, even if the behavior of each agent is determined by a Turing machine. Then, Roman turned to a variant of the model, where the agents are controlled by finite state machines. He demonstrated through a nicely illustrated proof that the demand of pheromones then increases to \Omega(D).

------------

Numerical Methods within the Ant Colony: The Illuminating Case of Multi-Objective Macronutrient Regulation in Eusocial Insects

In this work, Theodore Pavlic and Stephen Pratt study macronutrient regulation in ant colonies. Theodore presented results of measuring the intake of different macronutritions of ant colonies and argued that the nutrient regulation corresponds to a complex non-separable allocation problem. Furthermore, he presented evidence that the ants indeed manage to collaboratively solve this difficult problem efficiently.

Theodore also described their field experiments. As an example, they studied how the ants protect their food sources. They observed that  the protective behavior of a certain food source is independent of the demand of the dominant macronutrient in the corresponding food source.

-------------

Group discussion

A lengthy group discussion ensued the two talks. The main points of the discussions where how to create more incentives for people to apply for interdisciplinary PhD or postdoc positions and how to further develop already existing interdisciplinary positions. There seems to be a great demand for computer science (and in particular distributed computing) researchers in biology groups, since algorithmic and simulational challenges can often be handled more efficiently by computer scientists than biologists. The situation for biologists who want to work in CS seems a bit more difficult as pure biology questions do not tend to appear in CS too frequently. Moreover, biologists often lack the necessary fundamental CS education. Ideas to alleviate said problems mainly revolved around including more computer science aspects into biology curriculae and vice versa.