Decentralized Control and Interactive Design Methods for Large-Scale Heterogeneous Self-Organizing Swarms.
ABSTRACT We present new methods of decentralized control and interactive design for artificial swarms of a large number of agents that
can spontaneously organize and maintain non-trivial heterogeneous formations. Our model assumes no elaborate sensing, computation,
or communication capabilities for each agent; the self-organization is achieved solely by simple kinetic interactions among
agents. Specifications of the final formations are indirectly and implicitly woven into a list of different kinetic parameter
settings and their proportions, which would be hard to obtain with a conventional top-down design method but may be designed
heuristically through interactive design processes.
- SourceAvailable from: Hiroki Sayama
[Show abstract] [Hide abstract]
- "those old versions used a fixed number of swarms in each generation (though it was changeable manually in version 1.1). Moreover, the mutation operator became available only in version 1.1 (which was optional by default), and all the design results reported so far in ,  were produced using the mixing operator only. To address the problems mentioned above, we developed a new version of the Swarm Chemistry simulator, version 1.2. "
ABSTRACT: We developed Swarm Chemistry 1.2, a new version of the Swarm Chemistry simulator with an enhanced architecture of interactive evolutionary design for exploring heterogeneous self-propelled particle swarm dynamics. In the new version, each evolutionary operator acts locally and visually to part of the population of swarms on a screen, without causing entire generation changes that were used in earlier versions. This new architecture is intended to represent cognitive actions in human thinking and decision making processes more naturally. We tested the effectiveness of the new architecture through an in-class experiment with college students participating as designers as well as evaluators of swarms. We also measured the effects of mixing and mutation operators to the performance improvement of the design processes. The students' responses showed that the designs produced using the new version received significantly higher ratings from students than those produced using the old one, and also that each of the mixing and mutation operators contributed nearly independently to the improvement of the design quality. These results demonstrate the effectiveness of the new architecture of interactive evolutionary design, as well as the importance of having diverse options of action (i.e., multiple evolutionary operators in this context) in iterative design and decision making processes. This work also presents one of the few examples of human-involved experiments on the statistical evaluation of artificial lifeforms, whose quality (such as ldquolivingnessldquo) would be hard to assess without using human cognition at this point.Artificial Life, 2009. ALife '09. IEEE Symposium on; 05/2009
- "The evo-devo works of , , or with lesser morphogenetic abilities , , are among these few notable achievements. Other interesting studies have explored pairs of mechanisms: SA and PF, no GR—selfassembly based on cell adhesion and signaling pattern formation , but using only predefined cell types without internal genetic variables, e.g., ; PF and GR, no SA—non-trivial pattern formation from instruction-driven intercellular signaling , but on a fixed lattice without self-assembling motion, e.g., , ; SA and GR, no PF—heterogeneous swarms of genetically programmed, self-assembling particles, but in empty space without mutual differentiation signals, e.g., . Naturally, beyond the proof-of-concept simulations presented here, a more systematic exploration is needed. "
Conference Paper: Spatial Self-Organization of Heterogeneous, Modular Architectures.[Show abstract] [Hide abstract]
ABSTRACT: On the one hand, natural phenomena of spontane- ous pattern formation are generally random and repetitive, whereas, on the other hand, complicated heterogeneous architec- tures are the product of human design. The only examples of self- organized and structured systems are biological organisms pro- duced by development. Can we export their precise self-formation capabilities to computing systems? This work proposes an "em- bryomorphic engineering" approach inspired by evo-devo to solve the paradoxical challenge of planning autonomous systems. Its goal is to artificially reconstruct complex morphogenesis by integrating three fundamental ingredients: self-assembly and pattern formation under genetic regulation. It presents a spatial computational agent-based model that can be equivalently con- strued as (a) moving cellular automata, in which cell rearrange- ment is influenced by the pattern they form, or (b) heterogeneous swarm motion, in which agents differentiate into patterns accord- ing to their location. It offers a new abstract framework to ex- plore the causal and programmable link from genotype to pheno- type that is needed in many emerging computational domains, such as amorphous computing or artificial embryogeny.Second IEEE International Conference on Self-Adaptive and Self-Organizing Systems, SASO 2008, Workshops Proceedings, October 20-24, 2008, Venice, Italy; 01/2008
- [Show abstract] [Hide abstract]
ABSTRACT: Outside biological and social systems, natural pattern formation is essentially "simple" and random, whereas complicated struc-tures are the product of human design. So far, the only self-organized (undesigned) and complex morphologies that we know are biological organisms and some agent societies. Can we export their principles of decentralization, self-repair and evolution to our machines, networks and other artificial con-structions? In particular, can an "embryomorphic" engineering approach inspired by evo-devo solve the paradoxical challenge of planning autonomous systems? In this work, I wish to better understand and reproduce complex morphogenesis by investi-gating and combining its three fundamental ingredients: self-assembly and pattern formation under genetic regulation. The model I propose can be equivalently construed as (a) moving cellular automata, in which cell rearrangement is influenced by the pattern they form, or (b) heterogeneous collective motion, in which swarm agents differentiate into patterns according to their location. It offers a theoretical framework for exploring the causal and programmable link from genotype to phenotype.