Biomedical Engineering Reference
In-Depth Information
some extent contingent, nature of biological organization is in large part respon-
sible for this theoretical deficit.
There are, however, glimpses of intersection among principles—redun-
dancy, modularity, spatial compartmentalization, and distributed processing
share the use of a multiplicity of self-contained units discretely connected, to
ensure a degree of autonomy of processing. The feedback control, the develop-
mental module, and the connectionist model all exploit saturation effects to
damp the consequences of nonlinearity. Almost all the models assume some
form of sparse connectivity, whether it be among neurons, classes of mutation,
modules, signaling molecules, or immune effectors. From this perspective it can
be observed that robustness is a property of a large class of complex systems,
and that a general theory might be expressed in some abstract terms that tran-
scend system particulars. With a strong general theory of robustness we might
start exporting insights from evolved systems to engineered systems, where ro-
bustness is frequently minimal.
In a work in progress, we (33) have been developing a general theory of
robustness for energetic and informational flows over biological networks. This
work stresses the mechanisms by which networks, once perturbed, reconfigure
to alternative, redundant input sources so as to continue operating. This work
has also highlighted the fact that a satisfactory definition of robustness requires
some means of excluding inert or linear aggregates (systems of noninteracting
components) as robust. In other words, robustness theories need to include some
measure of contributions from both network flows (the system is in some sense
functional) and from invariance upon elimination of a subset of flows (the func-
tion can be retained upon perturbation). Robustness is a variational property of
evolved/engineered systems, and if we do not assume this dual measure ap-
proach (flows and invariance), rocks and dead organisms strike us as rather too
robust upon perturbation!
6.
REFERENCES
1. Aeyels D, Lamnabhi-Lagarrigue F, Schaft AJvd, Nonlinear Control Network. 1999.
Stability
and stabilization of nonlinear systems
. Springer, London.
2. Albert R, Jeong H, Barabasi AL. 2000. Error and attack tolerance of complex networks.
Nature
406
:378-382.
3. Alon U, Surette MG, Barkai N, Leibler S. 1999. Robustness in bacterial chemotaxis.
Nature
397
:168-171.
4. Ancel LW, Fontana W. 2000. Plasticity, evolvability, and modularity in RNA.
J Exp Zool
288
:242-283.
5. Barkai N, Leibler S. 1997. Robustness in simple biochemical networks.
Nature
387
:913-917.
6. Calabretta R, Nolfi S, Parisi D, Wagner GP. 1998. Emergence of functional modularity in
robots. In
From animals to animats
, pp. 497-504. Ed. R Pfeifer, B Blumberg, J-A Meyer, SW
Wilson. MIT Press, Cambridge.
