Catálogo de publicaciones - libros

This review paper sets out to explore the question of how future complex engineered systems based upon the swarm intelligence paradigm could be assured for dependability. The paper introduces the new concept of ‘swarm engineering’: a fusion of dependable systems engineering and swarm intelligence. The paper reviews the disciplines and processes conventionally employed to assure the dependability of conventional complex (and safety critical) systems in the light of swarm intelligence research and in so doing tries to map processes of analysis, design and test for safety-critical systems against relevant research in swarm intelligence. A case study of a swarm robotic system is used to illustrate this mapping. The paper concludes that while some of the tools needed to assure a swarm for dependability exist, many do not, and hence much work needs to be done before dependable swarms become a reality.

In this paper, we review methods used for macroscopic modeling and analyzing collective behavior of swarm robotic systems. Although the behavior of an individual robot in a swarm is often characterized by an important stochastic component, the collective behavior of swarms is statistically predictable and has often a simple probabilistic description. Indeed, we show that a class of mathematical models that describe the dynamics of collective behavior can be generated using the individual robot controller as modeling blueprint. We illustrate the macroscopic modelling methods with the help of a few sample results gathered in distributed manipulation experiments (collaborative stick pulling, foraging, aggregation). We compare the models’ predictions to results of probabilistic numeric and sensor-based simulations as well as experiments with real robots. Depending on the assumptions, the metric used, and the complexity of the models, we show that it is possible to achieve quantitatively correct predictions.

We consider swarms as systems with partial random synchronicity and look at the conditions for their convergence to a fixed point. The conditions turn out to be not much more stringent than for linear, one-step, stationary iterative schemes, either synchronous or sequential. The rate of convergence is also comparable. The main result is that swarms converge in cases when synchronous and/or sequential updating systems do not. The other significant result is that swarms can undergo a transition from non convergence to convergence as their degree of partial synchronicity diminishes, i.e., as they get more “disordered”. The production of order by disordered action appears as a basic characteristic of swarms.