LIVING SYSTEMS’ ORGANISATION AND PROCESSES FOR

ACHIEVING ADAPTATION

Principles to Borrow from Biology

Dragana Laketic and Gunnar Tufte

Department of Computer and Information Science, Norwegian University of Science and Technology

Sem Saelands vei 7-9, NO-7491 Trondheim, Norway

Keywords:

Adaptation, Homeostasis, Endocrine system, Hormones, Cybernetics, Living systems organisation.

Abstract:

Man–made systems, just like their biological counterparts, need to operate in a ﬂuctuating environment. Living

systems survive despite these ﬂuctuations. Their viability is made possible due to the ability to adapt to

environmental ﬂuctuations. Such ability the living systems possess is due to the organisation of these systems

and processes performed in response to ﬂuctuation. Therefore, deeper understanding of those aspects of

the living systems which make them adaptable may be beneﬁcial for human designers when faced with the

demands for the design of adaptive systems. This paper presents current state of our investigation and some

interesting postulations about how adaptation process may be sustained until adaptation is achieved in the

system under consideration. Further, we discuss some aspects of the living systems’ organisation which may

offer useful guidelines for adaptive systems design.

1 INTRODUCTION

No environment is static and all systems which are to

survive or maintain functionality, must adapt to the

changes and ﬂuctuations imposed. From the ﬁrst au-

tocatalytic cycles which exhibited basic characteris-

tics of life onwards, the living systems have been co-

evolving with their environment into more and more

complex systems characterised by particular organi-

sation. This organisation has enabled them to survive

and further procreate despite incessant environmental

changes and ﬂuctuations. The viability of the living

systems is therefore due to their organisation which

makes it possible for the systems to perform some

structural or functional change as a response to en-

vironmental ﬂuctuation so as to survive despite this

ﬂuctuation.

It is recognised that adaptations result from more

than one kind of adaptive processes which take place

at different levels. On a larger time–scale, co–

evolution of whole populations with their environ-

ments has produced living systems which are en-

dowed with inherent mechanisms for performing

adaptive processes. At individual level these mech-

anisms are activated once the ﬂuctuation in environ-

ment is detected so that the actual process of adapta-

tion is performed at a much smaller time–scale. This

coarse distinction related to the time scale at which

the adaptive processes occur, can serve as a basis for

further investigations into adaptability.

Adaptations are results of processes happening at

different hierarchical levels of the living systems, as

has been widely recognised and investigated (Belew

and Mitchell, 1996). At the level of the individ-

ual genotype, adaptations take place during evolu-

tionary processes or, to be more precise, during co-

evolutionary processes with the environment. Ac-

cording to natural selection laws, those individuals

which possess such genotypes which develop into the

ﬁttest phenotypes within the population environment,

will pass their genetic heritage to their offspring with

greatest probability. At the phenotypic level, differ-

ent forms of plasticity are exhibited by various sys-

tems, such as learning in neural system or certain be-

haviours, and they are all made possible thanks to

the organism’s organisation and accordingly inherent

adaptive mechanisms it possesses.

Out of the many facets of inherent mechanisms

in the living systems, processes which help organism

preserve its internal environment within the state of

equilibrium are prominent. This preservation, termed

homeostasis (Walter, 1967), is vital as it makes phys-

iological processes, the essence of life, to occur at the

right point in time and in the right order so that the life

254

Laketic D. and Tufte G. (2009).

LIVING SYSTEMS’ ORGANISATION AND PROCESSES FOR ACHIEVING ADAPTATION - Principles to Borrow from Biology.

In Proceedings of the International Joint Conference on Computational Intelligence, pages 254-259

DOI: 10.5220/0002335702540259

 SciTePress

is preserved. In keeping internal environment within

certain limits, a whole plethora of different processes

occur performed by a number of different systems in

a human body (Guyton and Hall, 2005). Such pro-

cesses exhibit remarkable characteristics with respect

to the usage of various mechanisms employed at dif-

ferent organisational levels. Not only are these mech-

anisms intertwined so as to achieve common goal –

adaptation, they also do so with the efﬁcient use of

resources.

The question naturally posed for a human designer

when faced with the task of designing a system to op-

erate in a ﬂuctuating environment, is: what can be

learned from living systems and what principles of

adaptiveness can be adopted in a man–made system?

This question has been asked by manyresearchers and

some of the results of such investigations are genetic

algorithms (Koza, 1992; Mitchell, 1999; de Jong,

2006), cybernetics (Wiener, 1948; Ashby, 1957) and,

if ’the ﬁrst notions of adaptation come from biology’

(Holland, 1992), then it can be said that any adapta-

tion in any man–made system is biologically inspired.

In this paper, we present the state of the ongo-

ing investigation into adaptation process inspired by

preservation of homeostasis and, in particular, en-

docrine system within homeostatic processes. Some

results regarding the adaptation process in a modular

system are commented in section 2. Further, direc-

tions for possible continuation of the work in this area

are discussed and some ideas for experiments are in-

troduced in sections 3 and 4. Finally, section 5 draws

a conclusion on the presented material.

2 HOMEOSTATIC PROCESSES

AND CONTROL AND

COMMUNICATION WITHIN

THEM

Preservation of homeostasis occurs thanks to the in-

terplay of several systems within the human body.

Within homeostatic processes, the role of communi-

cation and control is performed by endocrine system

and it can be said that homeostasis, as we know it to-

day, would be impossible without this system. En-

docrine system is responsible for secretion of hor-

mones, special substances which play the role of mes-

sengers. Hormones transmit information about the

change which has occurred in the body’s internal or

external environment. Hormones are secreted into

the bloodstream by special glands and tissues as a

response to some stimulus – a signal carrying infor-

mation about the change. They reach all the cells

via bloodstream and yet affect only the cells which

possess the matching receptor. The amounts of se-

creted hormones vary during their lifetime. In gen-

eral, these amounts are regulated through mechanisms

of positive and negative feedback. One of the remark-

able characteristics exhibited by hormones, is that the

small amounts of these substances can cause reac-

tions, or avalanche of reactions, which can have a

huge impact on the organism’s physiology and be-

haviour. In that sense, the hormones exhibit very efﬁ-

cient use of the resources available within the body.

The principles enabling the preservation of home-

ostasis have been studied by many researchers from

a more technical standpoint. In (Ashby, 1960), the

achievement of adaptation based on homeostatic prin-

ciples is related to the achievement of stability. There,

the environment is described through a set of param-

eters and the environmental ﬂuctuations are repre-

sented by the change in some of these parameters.

The state of the system is described by variables out

of which the set of essential variables is identiﬁed

whose preservation within certain limits is equivalent

to the preservation of homeostasis. The homeostatic

equilibrium within the system is, therefore, charac-

terised by the system ﬁnding itself in one of the ultra-

stable states.

Further, in (Neal and Timmis, 2003) artiﬁcial

homeostatic system is considered. Within such frame-

work, special attention has been paid to immune and

endocrine systems (Vargas et al., 2005; Moioli et al.,

2008), their mutual interaction and interaction(s) with

nervous system. Endocrine system has been further

studied for its many characteristics which are also

desirable in man–made systems. Examples come

mostly from the ﬁeld of evolutionary robotics (Paolo,

2003; Suzuki and Ikegami, 2008; Asada et al., 2008;

Moioli et al., 2009), but can be found also in self–

reconﬁgurable and swarm robotics (Shen et al., 2002)

and multiprocessor systems (Greensted and Tyrrell,

2005).

2.1 System under Investigation –

Architecture

In order to examine system adaptation to an environ-

mental ﬂuctuation, a model has been designed to meet

the simulation needs of such processes. Here we pro-

vide only its basic characteristics to enable easier un-

derstanding of the presented results for further dis-

cussion while details may be found in (Laketic et al.,

2009).

Figure 1 shows a schematic view of the model’s

architecture. It is modular where the modules – cells

are placed in a grid formation. Each cell is assumed

LIVING SYSTEMS' ORGANISATION AND PROCESSES FOR ACHIEVING ADAPTATION - Principles to Borrow

from Biology

255

Figure 1: Schematic view of the system’s architecture.

to possess a sensor which provides information about

the ﬂuctuation in some environmental parameter. In

this way, the measure of the ﬂuctuations in the cell’s

local environment is provided. It is further assumed

that each cell communicates via direct lines to its four

immediate neighbours to the north, east, south and

west. Such assumptions lend the model to the study

by the formalism of cellular automata (CA) (Sipper,

1997). Also, each cell possesses a set of tuning pa-

rameters which determine its functionality.

Figure 2: System conﬁguration under simulation and hor-

mone ﬂow loops between functionally related cells

The cell is identiﬁed by two kinds of identiﬁers,

similar to (Macias and Durbeck, 2004), one corre-

sponding to its physical placement in the grid and an-

other to its functional relatedness. Figure 2 shows one

example of such cell identiﬁcation. This conﬁgura-

tion was used in the experiments for which the results

are further presented. The cell marked with i is as-

sumed to receive functional input from the cell i− 1

and produce functional output for the cell i+ 1.

2.2 System Behaviour and Hormone

Flows Sustaining Adaptation

Hormone secretion is initiated dependent on the

sensed ﬂuctuation in the environmental parameter un-

der consideration. By the term hormone we assume

some kind of message which is transmitted around

the system architecture because of the sensed envi-

0

(L,Li,Ei)

4

(RP,Li,Ei)

3

(P,Li,Ei)

1

(S,Li,Ei+/-1)

9

(P,F,Ei+/-1)

2

(R,Li,Ei)

8

(L,F,Ei+/-1)

5

(SP,Li,Ei+/-1)

6

(S,A,Ei+/-1)

7

(SP,A,Ei+/-1)

4

()

0

(L,Li+/-1,Ei+/-1)

3

()

2

()

Figure 3: Finite state machine diagram which determines

the cell’s behaviour.

ronmental ﬂuctuation. The hormone secreted by the

cell which senses the ﬂuctuation is further transmitted

around the architecture in the manner the signals are

transmitted around CA. The cells are changing their

state according to the state diagram given in ﬁgure 3,

details being omitted for the sake of clarity. In this

way a hormone can reach all the cells within architec-

ture, as it does so via bloodstream in a living organ-

ism. In this case, it is only functionally related cells

which will recognise the incoming hormone, func-

tional relatedness, therefore, being analogous to the

receptors in their biological counterparts. It is further

assumed that adaptation process is conditioned by the

presence of the cell’s hormone and the hormone com-

ing from its functionally related cell.

The state of the system is monitored with respect

to the hormone ﬂow, functionality and environmen-

tal parameter under consideration. In (Laketic et al.,

2009) we have described how hormone secretion is

realised in the assumed architecture and investigated

the values for the parameters which determine the

hormone lifetime so that the presence of hormones is

ensured until adaptation is achieved.

Figure 4: System state machine diagram.

If the states of the system are considered in which

it can be while the cells are behaving according to the

state diagram presented in ﬁgure 3, it can be noticed

that the system may ﬁnd itself in one of the two states

– it is either adapting to new environmentalconditions

or it is adapted to these conditions. The case when the

system fails to adapt is disregarded for the reason that

IJCCI 2009 - International Joint Conference on Computational Intelligence

256

parameters which determine amounts of secreted hor-

mones are chosen so that the presence of hormones

is secured until adaptation is achieved, as presented

in (Laketic et al., 2009). Therefore, the system be-

haviour can be described by a state diagram as given

in ﬁgure 4.

When the system is in state 0, i.e. adapted to its

environment, all its cells are in either of the two states

marked with thicker borders in ﬁgure 3. These states

correspond to the cell being adapted to its local envi-

ronment. There, the values of the tuning parameters

are such that the cell performs desired functionality

under the local environmental condition.

Such view on the system’s behaviour draws anal-

ogy with the somewhat mechanistic approach to

the achievement of adaptation presented in (Ashby,

1960). When the system is in state 0, it is in one of

its ultrastable states, while the system state 1 refers to

the states through which the cells are passing while

performing adaptation process i.e. while rearranging

its tunable parameters so as to reach corresponding

ultrastable state thereby achieving adaptation.

Further, in (Laketic and Tufte, 2009) it has been

shown how formation of hormone ﬂow loops may

sustain adaptation processes until system adaptation

is achieved. There, the sensed ﬂuctuation in some en-

vironmentalparameter initiates secretion of hormones

whose recognition by functionally related cells fur-

ther initiates secretion of different types of hormones

thereby leading to the formation of the hormone ﬂow

loops. Such loops are closed between functionally re-

lated cells and, upon it, the adaptation process is ini-

tiated. It is sustained by the existing hormone ﬂow

loops until adaptation is achieved at the system level.

Figure 5: Hormone ﬂow loops which sustained adaptation

process until adaptation was achieved

Figure 5 shows the loops of hormone ﬂows which

are formed between functionally related cells. The

cells which sensed the ﬂuctuation in their local envi-

ronments were 52, 32, 23 – see ﬁgure 2, and the loops

were formed between them and their functionally re-

lated cells 53, 31, 22 respectivelyas shownin ﬁgure 5.

These loops sustained adaptation process until adap-

tation was achieved.

3 AN AUTONOMOUS AND

SELF–REFERRING UNIT OF A

LIVING SYSTEM

As said, the adaptiveness of living systems to envi-

ronmental ﬂuctuations is due to the inherent mech-

anisms they possess as a result of the co–evolution

with the environment. However, a more general ques-

tion of the organisation of living systems might pro-

vide a more general answer so as to offer more use-

ful principles for adaptive systems organisation. Or-

ganisation of living systems has become the focus of

the study of many researchers. Among the existing

theories, those which underline its autonomous and

self–referring nature are of particular interest when it

comes to addressing properties of adaptable system.

Investigation into the living systems organisation

is presented by the theories of the minimal system

which can be said to be alive. The model presented

in chemoton theory (Ganti et al., 2003) provides the

basic principles which deﬁne life. Within it, the

cyclic nature of the chemical reactions is recognised

and quantitatively studied so that the transition to the

emergence of life within such system can be distin-

guished. There, three different subsystems are recog-

nised according to the role they perform within the

system model. Similar ideas are presented in the

theory of hypercycles (Eigen and Schuster, 1979).

Again, the cyclic nature of reactions is prominent and

recognised at different hierarchical levels.

The theory presented in the seminal work on au-

topoiesis (Maturana and Varela, 1973) distinguishes

itself. The ensuing work pertaining to such Chilean

school of thought has resulted in the view on the liv-

ing systems’ organisation as an autonomous and self–

referring in its self–creation (therefore the term au-

topoiesis – ’self–creating’, coined by F. Varela to de-

note such organisation). These theories offer a frame-

work for studying dynamics of adaptive behaviour

and many models have been created to represent au-

tonomous and self–referring nature of the processes

taking place within such systems (McMullin, 1997;

Hutton, 2002).

It can be further investigated how systems of such

organisation adapt to environmental ﬂuctuations. In

order to do so, a suitable model for simulations would

be advantageous. The ﬁeld of artiﬁcial chemistries,

AChem, (Dittrich et al., 2008) offers the tools ﬂexible

enough to accommodate for this purpose, particularly

LIVING SYSTEMS' ORGANISATION AND PROCESSES FOR ACHIEVING ADAPTATION - Principles to Borrow

from Biology

257

when AChem models are applied for the information

processing task.

4 DISCUSSION

The results of our investigation into the aspects of

the organisation of living systems which provide their

adaptivenessto ﬂuctuating environment, have demon-

strated how a simpliﬁed control and communication

principles borrowed from endocrine system in the hu-

man body can be used to initiate and sustain adapta-

tion process until adaptation is achieved. The simula-

tions show the adaptation which is in agreement with

the theory of preservation of homeostasis as presented

in (Ashby, 1960). Moreover, we have postulated that

the loops formed by the hormone ﬂows can make up

the control part for adaptive mechanisms, as well as

realise communication within the assumed system ar-

chitecture. However, further investigation into the or-

ganisation at different hierarchical levels is expected

to provides solutions with the improved efﬁciency of

the adaptation process.

4.1 Adding Functionality

Present simulations do not consider which function-

ality is performed by a single cell or by a system. It

has only been assumed that some functionality is per-

formed and that it needs to be maintained despite en-

vironmental ﬂuctuations. Enhancement of the present

model with some functionality is advantageous for

the proof of principle of the existing ﬁndings. In

this respect, several AChem models have been ex-

amined and the current investigation revolves around

such simulations.

We begin with enhancingthe cell model within the

assumed architecture with some kind of metabolism.

The metabolism is affected by environmental ﬂuc-

tuations and so is the functionality within the cell.

The healthy metabolism is deﬁned and the goal of the

adaptation process is to keep the metabolism healthy

despite environmental ﬂuctuations. For the problem

at hand, such adaptive system can ensure the achieve-

ment of adaptation when the cell metabolism is re-

placed with the actual cell functionality.

4.2 Hierarchical Organisation

Presently considered model refers only to one hier-

archical level i.e. an architecture resembling tissue

formed of cells which follow the same rules of be-

haviour. It is our belief that further examination of

the presented control and communication principles

for adaptive processes at different hierarchical lev-

els of organisation, may result in ﬁndings which can

improve efﬁciency of achieving adaptation. The ef-

ﬁciency may be measured in time needed for the

system to achieve adaptation, the resources used to

achieve it or the complexity of the stages through

which the system achieves adaptation, if such stages

can be identiﬁed during adaptation process.

In the ﬁrst place, the formation of the tissue out of

the cells of the same type should be considered. Intro-

duction of different sets of rules which guide the cell’s

behaviour may determine different types of cells and

therefore different types of tissues. Further, organs

as units with some functionality may be assigned. Se-

cretion of control messengers can then be initiated at a

higher hierarchical level. Such models would further

allowfor the implementation of some of the endocrine

system characteristics which present model does not

support. Primarily, it refers to a whole avalanche

of reactions initiated by small amounts of hormones

once they reach the matching receptor.

5 CONCLUSIONS

This paper has presented some considerations regard-

ing the adaptiveness in biological systems and how

the underlying principles of this property of living

systems could be used for the design of adaptiveman–

made systems. Our investigation has tackled the prin-

ciples of homeostatic processes for the achievement

of adaptation within the assumed modular architec-

ture. In particular, endocrine system control and com-

munication role has been used to guide and sustain the

simulated adaptation process.

In the discussion section, some ideas have been

presented on what directions our future investiga-

tion may take. On one side, further development of

the model for simulations is considered with respect

to adding the functionality to the simulated system.

Such enhanced model is planned to serve for proving

the principle of the proposed adaptive technique. On

the other hand, hierarchical organisation of the living

systems need be further investigatedfor increasing the

efﬁciency of adaptation process.

Lastly, let us mention the challenge which remains

for the time beyond the completion of the investi-

gation – to design a unit in some technology which

would exhibit properties of the cell considered in our

simulations and with it lend itself to the construction

of an adaptive system based on the investigated adap-

tive principles borrowed from living systems. Sili-

con? Microﬂuids? Or, ....?

IJCCI 2009 - International Joint Conference on Computational Intelligence

258

REFERENCES

Asada, M., Hallam, J. C. T., Meyer, J.-A., and Tani,

J., editors (2008). Extended Homeostatic Adapta-

tion: Improving the Link between Internal and Be-

havioural Stability, Lecture Notes in Computer Sci-

ence. Springer.

Ashby, W. R. (1957). An Introduction to Cybernetics. Chap-

man & Hall Ltd.

Ashby, W. R. (1960). Design for a Brain, the origin of adap-

tive behaviour. Chapman & Hall Ltd.

Belew, R. K. and Mitchell, M. (1996). Adaptive Individu-

als in Evolving Populations: Models and Algorithms.

Addison-Wesley Publishing Company, Inc.

de Jong, K. (2006). Evolutionary Computation, a Uniﬁed

Approach. The MIT Press.

Dittrich, P., Ziegler, J., and Banzhaf, W. (2008). Artiﬁcial

chemistries – a review. In Artiﬁcial Life, pages 225–

275. Massachusetts Institute of Technology.

Eigen, M. and Schuster, P. (1979). The Hypercycle, a Prin-

ciple of Natural Self-Organisation. Springer-Verlag.

Ganti, T., (contributor), J. G., and (contributor), E. S.

(2003). The Principles of Life. Oxford University

Press.

Greensted, A. J. and Tyrrell, A. M. (2005). Implementation

results for a fault-tolerant multicellular architecture

inspired by endocrine communication. In NASA/DoD

Conference on Evolvable Hardware, Proceedings.

Guyton, A. and Hall, J. (2005). Textbook of Medical Physi-

ology. Elsevier Saunders.

Holland, J. H. (1992). Adaptation in Natural and Artiﬁcial

Systems. The MIT Press.

Hutton, T. (2002). Evolvable self–replicating molecules in

an artiﬁcial chemistry. In Artiﬁcial Life, pages 341–

356.

Koza, J. R. (1992). Genetic Programming: On the Pro-

gramming of Computers by Means of Natural Selec-

tion. MIT Press.

Laketic, D. and Tufte, G. (2009). Adaptation in tissue sus-

tained by hormonal loops. In 10th European Confer-

ence on Artiﬁcial Life. submitted.

Laketic, D., Tufte, G., and Haddow, P. (2009). Stochas-

tic adaptation to environmental changes supported by

endocrine system principles. In NASA/ESA Adaptive

Hardware and Systems, AHS 2009, accepted.

Macias, N. and Durbeck, L. (2004). Adaptive methods for

growing electronic circuits on an imperfect synthetic

matrix. In Bio Systems, pages 173–204. Elsevier.

Maturana, H. and Varela, F. (1973). Autopoiesis and Cog-

nition: the Realization of the Living. Dordecht: D.

Reidel Publishing Co.

McMullin, B. (1997). Computational autopoiesis: The orig-

inal algorithm. In Working Paper Number:97-01-001.

Mitchell, M. (1999). An Introduction to Genetic Algo-

rithms. The MIT Press.

Moioli, R., Vargas, P., Von Zuben, F., and Husbands, P.

(2008). Towards the evolution of an artiﬁcial home-

ostatic system. In IEEE Congress on Evolutionary

Computation 2008, Proceedings, pages 4023–4030.

Moioli, R. C., Vargas, P. A., and Husbands, P. (2009). A

multiple hormone approach to the homeostatic con-

trol of conﬂicting behaviours in an autonomous mo-

bile robot. In 2009 IEEE Congress on Evolutionary

Computation, Proceedings, pages 47–54.

Neal, M. and Timmis, J. (2003). Timidity: A useful emo-

tional mechanism for robot control? In Informat-

ica – special issue on perception and emotion based

control, pages 197–204. The Slovene Society Infor-

matika, Ljubljana, Slovenia.

Paolo, E. A. D. (2003). Organismically-inspired robotics:

Homeostatic adaptation and natural teleology beyond

the closed sensorimotor loop. In Dynamical sys-

tems approach to embodiment and sociality. Adelaide,

Australia: Advanced Knowledge International 2003,

pages 19–42.

Shen, W. M., Salemi, B., and Will, P. (2002). Hormone-

inspired adaptive communication and distributed con-

trol for conro self-reconﬁgurable robots. In IEEE

Transactions on Robotics and Automation, pages 700–

712.

Sipper, M. (1997). Evolution of Parallel Cellular Machines,

The Cellular Programming Approach. Springer.

Suzuki, K. and Ikegami, T. (2008). Homeodynamics in the

game of life. In Artiﬁcial Life XI: Proceedings of the

Eleventh International Conference on the Simulation

and Synthesis of Living Systems, pages 600–607. MIT

Press, Cambridge, MA.

Vargas, P., Moioli, R., Castro, L. N., Timmis, J., Neal, M.,

and Zuben, F. J. V. (2005). Artiﬁcial homeostatic sys-

tem: A novel approach. In European Conference in

Artiﬁcial Life, Lecture Notes in Artiﬁcial Intelligence,

pages 754–764. Springer-Verlag Berlin Heidelberg.

Walter, C. B. (1967). Wisdom of The Body. W.W.Norton

Wiener, N. (1948). Cybernetics or Control and Communi-

cation in the Animal and the Machine. The Technol-

ogy Press, Cambridge, Mass., John Wiley and Sons

Inc.

LIVING SYSTEMS' ORGANISATION AND PROCESSES FOR ACHIEVING ADAPTATION - Principles to Borrow

from Biology

259