Swarm Intelligence for Morphogenetic Engineering

Allen C. McBride; Bruce J. MacLennan

REVIEW 3 major objections 5 minor 45 references

Swarm Intelligence for Morphogenetic Engineering

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A morphogen-emission rule lets each agent read a global PDE specification from its local environment.

desk verdict The morphogen-based SPH compilation idea (Eq. 19) is mathematically neat and worth a serious referee, but the path-finding demo is qualitative and does not actually test the equilibrium assumptions underneath it. read the letter →

arxiv 1908.08787 v1 pith:NEBTLGID submitted 2019-08-13 cs.MA nlin.PSq-bio.TO

Bruce J. MacLennan , Allen C. McBride This is my paper · ORCID

classification cs.MAnlin.PSq-bio.TO

keywords activematterartificialmorphogenesismorphogeneticengineeringsmoothedparticlehydrodynamicsswarmintelligenceroboticsglobal-to-localcompilation

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper makes the case that developmental biology can supply a programming method for massive swarms of microscopic robots: describe the target structure as partial differential equations, then compile those global equations into local rules for individual agents. The authors present a PDE-based morphogenetic programming language, Morphgen, and demonstrate it on two bench problems: routing dense fiber bundles while avoiding obstacles, and assembling a segmented insect-like body with legs via a clock-and-wavefront segmentation process. Their central technical claim, developed last, is that an embodied form of smoothed particle hydrodynamics (SPH) solves the global-to-local compilation problem: if each agent emits a morphogen at rate $r_j = k m_j f(x_j)/\rho_j$, then at equilibrium the summed morphogen concentration sensed by an agent equals the SPH estimate of the desired field $f$, so the agent can read the global PDE locally. The payoff, if the claim holds, is a systematic code path from a continuum specification to individual agent control with no per-neighbor communication and with scalability across orders of magnitude in swarm size.

What carries the argument

The load-bearing mechanism is the production-rate identity Eq. (19), $r_j = k m_j f(x_j)/\rho_j$, which couples the global field specification to the local act of emitting a morphogen. Here $k$ is the morphogen decay rate, $m_j$ and $\rho_j$ are the mass and local density of agent $j$, and $f(x_j)$ is the desired field value carried by that agent. The identity makes the summed equilibrium concentration of the morphogen equal to an SPH estimate of $f$, and a rearrangement of the derivative-free SPH Laplacian (borrowed from the cited literature and reworked in Eq. 14) lets agents estimate diffusion terms from the same sensed field: $\langle \Delta f(x_i)\rangle = \frac{2}{\alpha}(\langle f(x_i)\rangle - f(x_i))$. The supporting object is the natural smoothing function (NSF): a physical field that an agent can generate, that falls off smoothly and monotonically, that sums additively with other agents' fields, and whose amplitude the agent can control; the morphogen version is the morphogen smoothing function (MSF). The NSF replaces explicit communication, so each agent's sensing operation is $O(1)$.

What would settle it

In a physical testbed or a high-fidelity simulation with moving, finite-size agents, measure the concentration profile around a single agent moving at constant speed and compare the gradient sensed by that agent with the gradient of the idealized morphogen smoothing function in Table 1; if the calibrated residual grows with speed or agent diameter, then the equality between sensed field and SPH estimate fails and the compilation claim breaks. A simpler test is to run the Sec. 6.9 path-finding program with motion-distortion calibration disabled and observe whether path quality degrades monotonically with agent speed or size.

Watch

Extended reading notes

Core claim

At the center of the paper is the observation that the standard SPH field estimate $\langle f(x)\rangle = \sum_j \frac{m_j}{\rho_j} f(x_j) W_j(x)$ need not be computed by gathering messages from neighbors; it can be built into the physics of the environment. If agent $j$ emits a diffusible morphogen whose equilibrium concentration profile is proportional to the smoothing kernel $W_j$, and if it controls its emission rate according to $r_j = k m_j f(x_j)/\rho_j$ (Eq. 19), then the contribution of agent $j$ to the sensed concentration at another point is exactly $\frac{m_j}{\rho_j} f(x_j) W_j(x)$. Since morphogen concentrations from many agents add in the medium, the total local concentration is the SPH estimate of $f$, and an agent sensing the value and gradient of that concentration obtains the field and its derivatives with $O(1)$ sensing rather than $O(n)$ communication. The paper names such physical fields natural smoothing functions, and the morphogen-based special case morphogen smoothing functions, giving idealized one-, two-, and three-dimensional profiles (Table 1) under equilibrium diffusion-decay. Its Sec. 6.9 path-finding simulation, run with agent counts spanning a factor of about 64, is presented as evidence that this compilation works for a Morphgen program whose specified fields have nothing to do with fluids.

Load-bearing premise

The claim stands or falls on whether agents can be treated as stationary point sources in a medium where morphogens obey linear diffusion with uniform decay and are at equilibrium with production, because under those conditions the sensed concentration equals the field estimate and the paper's corrections for moving, finite-size agents are only approximate.

Editorial extensions

If this is right

Any Morphgen program whose non-swarm fields can be treated as agent-carried state compiles, in principle, to swarm behavior by associating each field with a physical morphogen; no explicit neighbor-to-neighbor data links are required.
Because the continuum specification is independent of agent size, the same morphogenetic program should remain correct as agents shrink and multiply, with the paper demonstrating qualitative similarity across a 64-fold change in agent count in the path-finding example.
A small library of physical morphogens with fixed diffusion and decay rates can be reused to implement different abstract fields, since the production-rate identity specifies the correct emission rate without tuning the smoothing shape.
SPH swarm control is extended beyond fluid-like swarms to arbitrary PDE-specified fields, such as velocity-based flocking and path-planning cue fields, not just pressure-like physics quantities.
The clock-and-wavefront example shows that the same morphogenetic motif can be redeployed for different structures (spine and legs), suggesting that these PDE-level processes are composable programming primitives.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Editorial inference: because the natural smoothing function requirements do not mention chemistry, the same compilation scheme could be implemented with acoustic or electromagnetic fields distinguished by frequency, which would remove diffusion lag and moving-agent distortion and make the method usable in dry environments.
Editorial inference: the transient correction for agent motion is a calibration table rather than a closed-loop control, so a natural testable extension is a controller that actively maintains equilibrium, for example by advecting the morphogen field opposite to agent motion, and checking whether the calibrated residual shrinks.
Editorial inference: if the emission rule were implemented in engineered microorganisms that both produce and sense signaling molecules, the approach suggests a low-power, wireless-free route to collective computation at cellular scales, with the morphogen field doubling as communication and structural material.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

The paper's real asset is the middle of Sec. 6: the natural smoothing function idea and the derivation of Eq. 19, r_j = k m_j f(x_j)/rho_j, which ties an agent's morphogen emission rate to the SPH estimate of a desired field. That is a compact, clean result, and the math checks out under the stated idealizations (point sources, Fickian diffusion, uniform decay, unmoving agents, equilibrium). The MSF table with Bessel-function kernels is a nice touch. This genuinely goes beyond the usual SPH-swarm literature, which mostly emulates fluids; here the SPH machinery is repurposed for arbitrary PDEs specified in their Morphgen language.

The two earlier examples (neural fiber routing and the clock-and-wavefront spine/leg segmentation) are detailed enough to reimplement, and they illustrate the language, but they are not the main contribution.

The soft spots are real and mostly in the validation. The path-finding demo in Sec. 6.9 is qualitative: Table 2 shows snapshots at three scales with no error bars, no comparison of sensed fields to the SPH/PDE targets, and no sensitivity analysis. More importantly, the demo uses moving, disc-shaped agents, while Eq. 19 is derived for unmoving point sources. The authors acknowledge this in Sec. 6.7 and offer only approximate corrections. So the demo does not actually test the equilibrium mechanism under its own assumptions; it shows behavior that looks path-like, but not that the MSF field matches the intended field. The stress-test note's concern is valid: the one proposed validation doesn't isolate Eq. 19's correctness.

There is also a constructive, not predictive, character to the formalism: agents are told to emit exactly enough morphogen to make concentration match a target field, so the examples demonstrate consistency rather than test a prediction. That is common for design frameworks, but it lowers the evidentiary weight.

Still, these are weaknesses in validation, not in the core math. The paper is honest about its limitations—Sec. 6.7 and 6.8 concede the transient and continuity issues explicitly—and the derivations are transparent. I think the reader's CONDITIONAL verdict is about right: the paper deserves a serious referee, but not a clean accept as is. I'd want code/data and a quantitative check of Eq. 19 in a controlled static scenario before believing the compilation claim.

Who benefits: people working on programmable matter, morphogenetic engineering, and SPH-based swarm control. It is a good discussion paper, but treat the validation as preliminary.

My recommendation: send to peer review, but ask the authors to release code and add a quantitative static-agent test of Eq. 19.

Referee Report

3 major / 5 minor

Summary. This paper argues that embryological morphogenesis offers a model for coordinating massive swarms to assemble complex, multiscale structures, and it develops a PDE-based morphogenetic programming language (Morphgen) with a prototype implementation. Two application examples are presented: routing dense neural-fiber bundles via modified flocking (Sec. 4) and assembling a segmented insect-like robot body using a clock-and-wavefront process (Sec. 5). The paper's main technical contribution, in Sec. 6, is an embodied variation of smoothed particle hydrodynamics (SPH) for global-to-local compilation: under assumptions of point, unmoving agents in diffusion-decay equilibrium, the production rate r_j = k m_j f(x_j)/rho_j (Eq. 19) is derived so that summed morphogen concentration equals the SPH estimate of an arbitrary field f. Corrections for transient and motion-induced effects are proposed in Sec. 6.7, and a multi-scale path-finding simulation (Table 2) is offered as an illustration.

Significance. If the compilation mechanism were quantitatively validated, the NSF/MSF framework would be a noteworthy advance: it replaces O(n) neighbor-to-neighbor communication in traditional SPH swarm control with O(1) local sensing of a natural physical field, and it extends SPH-based robotics from fluid emulation toward arbitrary Morphgen PDEs. The derivation of Eq. (19) is internally consistent under its stated assumptions, the paper is explicit about the limitations of its transient corrections, and the multi-scale path-finding demo is honest about being qualitative. These strengths are real and should be credited. The central weakness is evidentiary: the only validation of the load-bearing Eq. (19) is a qualitative simulation that violates the derivation's assumptions, and no quantitative error analysis is reported.

major comments (3)

[Sec. 6.6, Eq. (19); Sec. 6.9, Table 2] The central claim that Eq. (19) enables faithful global-to-local compilation is not supported by the presented evidence. The derivation assumes point sources, unmoving agents, and production-diffusion-degradation equilibrium (Sec. 6.6), while the path-finding simulation uses disk-shaped agents moving through a flowing medium and explicitly includes transient and motion-induced distortions (Sec. 6.7). No result in Table 2 or elsewhere reports a quantitative comparison between the sensed morphogen field (or the resulting density field) and the target SPH estimate or PDE solution. Since the paper argues the example validates the compilation mechanism, the absence of such a comparison is a load-bearing gap.
[Sec. 6.7.2] The correction for motion-induced distortion is acknowledged to be partial and implementation-dependent: calibration is proposed only for an agent's own contribution to its gradient estimates, while neighbor contributions and curved paths are explicitly left uncorrected. The text does not bound the residual error or quantify how far the sensed field deviates from the equilibrium SPH estimate for the simulation's agent speeds and sizes. Without such an error analysis, the simulation cannot establish that Eq. (19)'s equilibrium relation holds approximately under the conditions of the demo.
[Sec. 6.4, Eq. (14)] The derivation of the SPH Laplacian replaces the sum \sum_j (m_j/rho_j) W_j(x_i) by 1, treating the SPH estimate of unity as exact. For finite neighbor counts and non-uniform agent distributions this is an approximation whose magnitude is not bounded or discussed. This matters because the path-finding program (Sec. 6.9) uses Laplacian terms (e.g., kappa_S3 del^2 S) that would be computed through this formula. The paper should either provide a bound, cite standard SPH error estimates, or explicitly note the limitation as a source of error in the proposed compilation pipeline.

minor comments (5)

[Sec. 3.3] The citation '[?]p. 20]Fleischer-PhD' is an unresolved reference and should be replaced with the proper bibliographic entry.
[References] Several references are incomplete: [8] lacks volume/pages, [10] lacks page numbers, [12], [15], [19], and [37] lack journal names or full bibliographic details, and [44] lacks article numbering. The reference list needs a careful final edit.
[Table 1] In the one-dimensional row, the normalized profile sqrt(k/E) e^{-sqrt(k/E) r} integrates to 1 over the half-line r > 0, but for a point source in an unbounded one-dimensional medium the symmetric two-sided solution has coefficient sqrt(k/E)/2; the table should specify which domain is intended.
[Sec. 4.2, Eq. (2)] Equation (2) contains an unmatched closing bracket: "V = v[(1-lambda)V1 - lambda V2 + kW DW 2)]" has one more ')' than '('.
[Figure 4] The panels (a)-(c) do not label axes or provide color/scale bars, so the reader cannot infer quantitative concentration or density values from the displayed fields.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the SPH morphogen production law is an explicitly solved design equation, not a fitted prediction; the simulation shortfall is a validation gap, not a circular reduction.

full rationale

The claimed derivation chain in Sec. 6.6 is self-contained and constructive. Eq. (15) states the design goal (an SPH estimate equals the summed morphogen concentration), Eq. (16) decomposes that goal per agent, and Eq. (19) is obtained by algebraically solving for the emission rate using the equilibrium relation C_j = r_j/k. This is an inverse control law, not a quantity fitted to data and then relabeled as a prediction; the equality is by construction only in the legitimate sense that a compiler is allowed to choose controls that realize its specification. The paper does not use measured morphogen concentrations to set r_j and then claim to predict those concentrations. The moving, disk-shaped agents and transient fields of Sec. 6.9 do not satisfy the point-source, equilibrium, unmoving-agent assumptions of Sec. 6.6, and the simulation is qualitative, but that is an unvalidated-application concern, not circularity. The segment counts in Sec. 5.4 are also chosen by setting the timer G_S from the desired N_S, again an explicit design parameter, not a predicted outcome. Self-citations [20]-[30] support the language and earlier examples, but the SPH/NSF derivation itself does not reduce to those citations and no uniqueness theorem or ansatz is imported from them. No circular step meeting the quoted-equation standard was found.

Assumptions & free parameters 12 free parameters · 6 assumptions · 0 invented entities

No new physical entities (forces, new particles, or new dimensions) are postulated. The 'natural smoothing function' and 'morphogen smoothing function' are conceptual tools, and variables such as G, L, and phi are internal model state rather than physical entities. The central claims rest on many hand-selected parameters and on the continuum and equilibrium assumptions, which are acknowledged but not independently validated.

free parameters (12)

dA = 0.03
Attractant diffusion constant in path routing (Sec. 4.2); no sensitivity analysis.
tauA = 100
Attractant decay time constant in path routing (Sec. 4.2).
kG = 100
Goal material attractant release rate (Sec. 4.2).
aP = 20
Autocatalytic rate of path material (Eq. 1).
thetaP = 0.3
Path autocatalysis threshold (Eq. 1).
lambda = 0.1
Density regularization weight in swarm velocity (Eq. 2).
kW = 0.3
Random motion coefficient in swarm velocity (Eq. 2).
tD = 5
Time delay before swarm movement (Sec. 4.2).
kP = 30
Path deposition rate (Sec. 4.2).
tauP = 0.2
Attractant absorption time constant by path material (Sec. 4.2).
nuS = 1/(2*pi) or 1/pi
Pacemaker frequency for spinal segmentation; varied in Fig. 9.
GS = formula to produce NS segments
Initial growth timer set by GS = thetaG exp(NS/(nuS tauG)) to force a chosen segment count (Sec. 5.4).

assumptions (6)

domain assumption Agent swarms can be treated as continuous density fields (continuum approximation).
Used throughout Secs. 2-5 to justify PDE descriptions; valid only for very large numbers of very small agents.
domain assumption Morphogens obey Fickian diffusion and uniform first-order decay in a homogeneous medium.
Assumed in Sec. 6.6 for MSF derivation and in all PDE examples.
domain assumption Morphogen fields superpose linearly at an agent's sensors.
Required for NSF/MSF framework (Secs. 6.3-6.4) so that summed concentrations act as SPH estimates.
standard math SPH smoothing functions have unity integral, are symmetric, and decrease smoothly and integrably.
Standard SPH requirements cited to ref [18] and used to derive production rates (Sec. 6.3, Eq. 19).
domain assumption For MSF equilibrium, agents are points, unmoving, and production, diffusion, and degradation are at steady state.
Stated in Sec. 6.6; later corrected only approximately (Sec. 6.7).
domain assumption The clock-and-wavefront model is an appropriate template for artificial segmentation.
The design of Sec. 5 copies the biological model and assumes it will be robust in engineered swarms.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Swarm Intelligence for Morphogenetic Engineering." pith.science (2026). https://pith.science/paper/NEBTLGID

@misc{pith2026190808787,
  author       = {Pith},
  title        = {Pith review of: Swarm Intelligence for Morphogenetic Engineering},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NEBTLGID}},
  note         = {Machine review of arXiv:1908.08787}
}

read the original abstract

We argue that embryological morphogenesis provides a model of how massive swarms of microscopic agents can be coordinated to assemble complex, multiscale hierarchical structures. This is accomplished by understanding natural morphogenetic processes in mathematical terms, abstracting from the biological specifics, and implementing these mathematical principles in artificial systems. We have developed a notation based on partial differential equations for artificial morphogenesis and have designed a prototype morphogenetic programming language, which permits precise description of morphogenetic algorithms and their automatic translation to simulation software. Morphogenetic programming is illustrated by two examples: (1) use of a modified flocking algorithm to route dense fiber bundles between regions of an artificial cortex while avoiding other bundles; (2) use of the clock-and-wavefront model of spinal segmentation for the assembly of the segmented spine of an insect-like robot body and for assembling segmented legs on the robot's spine. Finally, we show how a variation of smoothed particle hydrodynamics (SPH) swarm robotic control can be applied to the global-to-local compilation problem, that is, the derivation of individual agent control from global PDE specifications.

Figures

Figures reproduced from arXiv: 1908.08787 by the authors.

**Figure 2.** Simulations of neural path formation. In both simulations it can be seen that a swarm sometimes splits to go [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗

**Figure 3.** Simulation of swarm assembly of insect-like robot body. The body is growing toward the right; the head [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗

Figures from the paper (6 more)

**Figure 4.** Figure 4: Simulation of spine growth process in progress ( [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]

**Figure 5.** Figure 5: Diffusion of C and R morphogens after seven new spinal segments have differentiated (tan color). Blue color at far right represents C (caudal morphogen) diffusing from the tailbud. Green color around the segments represents R (rostral morphogen) diffusing from differen…

**Figure 6.** Figure 6: Propagation of α morphogen. Two successive waves are propagating leftward toward the head. The decaying α pulse in the tailbud is visible at the righthand end. 15 [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]

**Figure 7.** Figure 7: Three-dimensional simulation of segmentation. Three new segments posterior to the head have been assem [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]

**Figure 8.** Figure 8: Diffusion of a and p morphogens from anterior and posterior tissue, respectively, in each differentiated segment. Orange color represents a morphogen diffusing from differentiated anterior tissue, and green color represents p morphogen diffusing from differentiated pos…

**Figure 9.** Figure 9: Segmentation with higher pacemaker frequency. This simulation has twice the pacemaker frequency [PITH_FULL_IMAGE:figures/full_fig_p018_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 45 canonical work pages

[1]

Knight, Jr., Radhika Nag- pal, Erik Rauch, Gerald Jay Sussman, and Ron Weiss

Harold Abelson, Don Allen, Daniel Coore, Chris Hanson, George Homsy, Thomas F. Knight, Jr., Radhika Nag- pal, Erik Rauch, Gerald Jay Sussman, and Ron Weiss. Amorphous computing. Commun. ACM, 43(5):74–82, May 2000

work page 2000
[2]

Bourgine and A

P. Bourgine and A. Lesne, editors. Morphogenesis: Origins of Patterns and Shapes, Berlin, 2011. Springer

work page 2011
[3]

Cooke and E

J. Cooke and E. C. Zeeman. A clock and wavefront model for control of the number of repeated structures during animal morphogenesis. J. Theoretical Biology, 58:455–76, 1976

work page 1976
[4]

P. G. de Gennes. Soft matter. Science, 256:495–497, 1992

work page 1992
[5]

Dequ ´eant and O

M.-L. Dequ ´eant and O. Pourqui ´e. Segmental patterning of the vertebrate embryonic axis. Nature Reviews Genetics, 9:370–82, 2008

work page 2008
[6]

Yeomans, and Francesc Sagu´es

Amin Doostmohammadi, Jordi Ign ´es-Mullol, Julia M. Yeomans, and Francesc Sagu´es. Active nematics. Nature Communications, 9(1):3246, 2018

work page 2018
[7]

Organically grown architectures: Creating decentralized, autonomous systems by embryomorphic engineering

Ren ´e Doursat. Organically grown architectures: Creating decentralized, autonomous systems by embryomorphic engineering. In R. P. W¨urtz, editor, Organic Computing, pages 167–200. Springer, 2008

work page 2008
[8]

A review of morphogenetic engineering

Ren ´e Doursat, Hiroki Sayama, and Olivier Michel. A review of morphogenetic engineering. 12(4):517–535

work page

Show all 45 references

[9]

Gabor Forgacs and Stuart A. Newman. Biological Physics of the Developing Embryo . Cambridge University Press, Cambridge, UK, 2005

2005
[10]

Self-swarming robots that exploit hydrodynamical interaction

Ryo Fujiwara, Takeshi Kano, and Akio Ishiguro. Self-swarming robots that exploit hydrodynamical interaction. pages 1–7
[11]

Giavitto and A

J.-L. Giavitto and A. Spicher. Computer morphogenesis. In P. Bourgine and A. Lesne, editors, Morphogenesis: Origins of Patterns and Shapes, pages 315–340. Springer, Berlin, 2011

2011
[12]

Gingold and Joseph J

Robert A. Gingold and Joseph J. Monaghan. Smoothed particle hydrodynamics: theory and application to non- spherical stars. 181(3):375–389
[13]

S. C. Goldstein, J. D. Campbell, and T. C. Mowry. Programmable matter. Computer, 38(6):99–101, June 2005

2005
[14]

Space-Time Continuous Models of Swarm Robotic Systems , volume 9 of Cognitive Systems Monographs

Heiko Hamann. Space-Time Continuous Models of Swarm Robotic Systems , volume 9 of Cognitive Systems Monographs. Springer Berlin Heidelberg
[15]

Huang, J

C. Huang, J. M. Lei, M. B. Liu, and X. Y . Peng. An improved KGF-SPH with a novel discrete scheme of laplacian operator for viscous incompressible ﬂuid ﬂows. 81(6):377–396
[16]

H. Kitano. Morphogenesis for evolvable systems. In E. Sanchez and M. Tomassini, editors, Towards Evolvable Hardware: The Evolutionary Engineering Approach, pages 99–117. Springer, Berlin, 1996

1996
[17]

Lichtman and Winfried Denk

Jeff W. Lichtman and Winfried Denk. The big and the small: Challenges of imaging the brain’s circuits.Science, 334(6056):618–623, 4 November 2011

2011
[18]

G. R. Liu and M. B. Liu. Smoothed particle hydrodynamics: a meshfree particle method . World Scientiﬁc. OCLC: ocm52947194
[19]

Leon B. Lucy. A numerical approach to the testing of the ﬁssion hypothesis. 82:1013–1024
[20]

MacLennan

Bruce J. MacLennan. Preliminary development of a formalism for embodied computation and morphogenesis. Technical Report UT-CS-09-644, Department of Electrical Engineering and Computer Science, University of Tennessee, Knoxville, TN, 2009

2009
[21]

MacLennan

Bruce J. MacLennan. Models and mechanisms for artiﬁcial morphogenesis. In F. Peper, H. Umeo, N. Matsui, and T. Isokawa, editors, Natural Computing, Springer series, Proceedings in Information and Communications Technology (PICT) 2, pages 23–33, Tokyo, 2010. Springer

2010
[22]

MacLennan

Bruce J. MacLennan. Morphogenesis as a model for nano communication. Nano Communication Networks. , 1(3):199–208, 2010

2010
[23]

MacLennan

Bruce J. MacLennan. Artiﬁcial morphogenesis as an example of embodied computation. International Journal of Unconventional Computing., 7(1–2):3–23, 2011

2011
[24]

MacLennan

Bruce J. MacLennan. Embodied computation: Applying the physics of computation to artiﬁcial morphogenesis. Parallel Processing Letters, 22(3):1240013, 2012

2012
[25]

MacLennan

Bruce J. MacLennan. Molecular coordination of hierarchical self-assembly. Nano Communication Networks , 3(2):116–128, June 2012. 30 A PREPRINT - AUGUST 26, 2019

2012
[26]

MacLennan

Bruce J. MacLennan. Coordinating massive robot swarms. International Journal of Robotics Applications and Technologies, 2(2):1–19, 2014

2014
[27]

MacLennan

Bruce J. MacLennan. The morphogenetic path to programmable matter. Proceedings of the IEEE, 103(7):1226– 1232, 2015

2015
[28]

MacLennan

Bruce J. MacLennan. Coordinating swarms of microscopic agents to assemble complex structures. In Ying Tan, editor, Swarm Intelligence, Vol. 1: Principles, Current Algorithms and Methods , PBCE 119, chapter 20, pages 583–612. Institution of Engineering and Technology, 2018

2018
[29]

MacLennan

Bruce J. MacLennan. The Synmac syntax macroprocessor: Introduction and manual, version 5. Faculty Publi- cations and Other Works — EECS. http://trace.tennessee.edu/utk elecpubs/23, University of Tennessee Depart- ment of Electrical Engineering and Computer Science, 2018

2018
[30]

MacLennan

Bruce J. MacLennan. A morphogenetic program for path formation by continuous ﬂocking. International Journal of Unconventional Computing, 14:91–119, 2019

2019
[31]

Maningo, Gerard Ely U

Jose Martin Z. Maningo, Gerard Ely U. Faelden, Reiichiro Christian S. Nakano, Argel A. Bandala, Ryan Rhay P. Vicerra, and Elmer P. Dadios. Formation control in quadrotor swarm aggregation using smoothed particle hy- drodynamics. In Region 10 Conference (TENCON), 2016 IEEE, pag...

2016
[32]

Murata and H

S. Murata and H. Kurokawa. Self-reconﬁgurable robots: Shape-changing cellular robots can exceed conventional robot ﬂexibility. IEEE Robotics & Automation Magazine, pages 71–78, March 2007

2007
[33]

Programming methodology for biologically-inspired self-assembling systems

Radhika Nagpal, Attila Kondacs, and Catherine Chang. Programming methodology for biologically-inspired self-assembling systems. In AAAI Spring Symposium on Computational Synthesis: From Basic Building Blocks to High Level Functionality, March 2003

2003
[34]

Active matter at the interface between materials science and cell biology

Daniel Needleman and Zvonimir Dogic. Active matter at the interface between materials science and cell biology. Nature Reviews Materials, 2:17048 EP –, 07 2017

2017
[35]

Bio-inspired self-organising multi-robot pattern formation: A review

Hyondong Oh, Ataollah Ramezan Shirazi, Chaoli Sun, and Yaochu Jin. Bio-inspired self-organising multi-robot pattern formation: A review. 91:83–100
[36]

Pac, Aydan M

Muhammed R. Pac, Aydan M. Erkmen, and Ismet Erkmen. Control of robotic swarm behaviors based on smoothed particle hydrodynamics. In Intelligent Robots and Systems, 2007. IROS 2007. IEEE/RSJ International Conference on, pages 4194–4200. IEEE

2007
[37]

Perkinson and Bahram Shafai

James R. Perkinson and Bahram Shafai. A decentralized control algorithm for scalable robotic swarms based on mesh-free particle hydrodynamics. In Proc. of the IASTED Int. Conf. on Robot. and Applications , pages 102–107, 2005

2005
[38]

Luciano C. A. Pimenta, Guilherme A. S. Pereira, Nathan Michael, Renato C. Mesquita, Mateus M. Bosque, Luiz Chaimowicz, and Vijay Kumar. Swarm Coordination Based on Smoothed Particle Hydrodynamics Technique. IEEE Transactions on Robotics, 29(2):383–399, April 2013

2013
[39]

Salazar-Ciudad, J

I. Salazar-Ciudad, J. Jernvall, and S. A. Newman. Mechanisms of pattern formation in development and evolu- tion. Development, 130:2027–37, 2003

2027
[40]

Silic, Zhuoyuan Song, and Kamran Mohseni

Matthew B. Silic, Zhuoyuan Song, and Kamran Mohseni. Anisotropic ﬂocking control of distributed multi-agent systems using ﬂuid abstraction. In 2018 AIAA Information Systems-AIAA Infotech @ Aerospace . American Institute of Aeronautics and Astronautics

2018
[41]

Spicher, Olivier Michel, and J.-L

A. Spicher, Olivier Michel, and J.-L. Giavitto. Algorithmic self-assembly by accretion and by carving in MGS. In Proc. of the 7th International Conference on Artiﬁcial Evolution (EA ‘05), number 3871 in Lecture Notes in Computer Science, pages 189–200. Springer-Verlag, Berlin, 2005

2005
[42]

Larry A. Taber. Nonlinear Theory of Elasticity: Applications in Biomechanics . World Scientiﬁc, Singapore, 2004

2004
[43]

Teague, Patrick Guye, and Ron Weiss

Brian P. Teague, Patrick Guye, and Ron Weiss. Synthetic morphogenesis. page a023929
[44]

Tilki and A

U. Tilki and A. M. Erkmen. Fluid swarm formation control for hand gesture imitation by ellipse ﬁtting. 51(6):473–475
[45]

Material approaches to active tissue mechanics

Wang Xi, Thuan Beng Saw, Delphine Delacour, Chwee Teck Lim, and Benoit Ladoux. Material approaches to active tissue mechanics. Nature Reviews Materials, 4(1):23–44, 2019. 31

2019

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Knight, Jr., Radhika Nag- pal, Erik Rauch, Gerald Jay Sussman, and Ron Weiss

Harold Abelson, Don Allen, Daniel Coore, Chris Hanson, George Homsy, Thomas F. Knight, Jr., Radhika Nag- pal, Erik Rauch, Gerald Jay Sussman, and Ron Weiss. Amorphous computing. Commun. ACM, 43(5):74–82, May 2000

work page 2000

[2] [2]

Bourgine and A

P. Bourgine and A. Lesne, editors. Morphogenesis: Origins of Patterns and Shapes, Berlin, 2011. Springer

work page 2011

[3] [3]

Cooke and E

J. Cooke and E. C. Zeeman. A clock and wavefront model for control of the number of repeated structures during animal morphogenesis. J. Theoretical Biology, 58:455–76, 1976

work page 1976

[4] [4]

P. G. de Gennes. Soft matter. Science, 256:495–497, 1992

work page 1992

[5] [5]

Dequ ´eant and O

M.-L. Dequ ´eant and O. Pourqui ´e. Segmental patterning of the vertebrate embryonic axis. Nature Reviews Genetics, 9:370–82, 2008

work page 2008

[6] [6]

Yeomans, and Francesc Sagu´es

Amin Doostmohammadi, Jordi Ign ´es-Mullol, Julia M. Yeomans, and Francesc Sagu´es. Active nematics. Nature Communications, 9(1):3246, 2018

work page 2018

[7] [7]

Organically grown architectures: Creating decentralized, autonomous systems by embryomorphic engineering

Ren ´e Doursat. Organically grown architectures: Creating decentralized, autonomous systems by embryomorphic engineering. In R. P. W¨urtz, editor, Organic Computing, pages 167–200. Springer, 2008

work page 2008

[8] [8]

A review of morphogenetic engineering

Ren ´e Doursat, Hiroki Sayama, and Olivier Michel. A review of morphogenetic engineering. 12(4):517–535

work page