Papers by George Mitchell
European Conference on Complex Systems, 2008
We present a novel cost benefit operator that assists multi level genetic algorithm searches. Thr... more We present a novel cost benefit operator that assists multi level genetic algorithm searches. Through the use of the cost benefit operator, it is possible to dynamically constrain the search of the base level genetic algorithms, to suit the user's requirements. We note that the current literature has abundant studies on meta evolutionary GA's, however these approaches have not identified an efficient approach to the termination of base GA searchs or a means to balance practical consideration such as quality of solution and the expense of computation. Our Quality time tradeoff operator (QTT) is user defined, and acts as a base level termination operator and also provides a fitness value for the meta-level GA. In this manner, the amount of computation time spent on less encouraging configurations can be specified by the user. Our approach was applied to a computationally intensive test problem which evaluates a large set of configuration settings for the base GAs to find suitable configuration settings (population size, crossover operator and rate, mutation operator and rate, repair or penalty and the use of adaptive mutation rates) for selected TSP problems.
European Conference on Complex Systems, Sep 1, 2006
We are investigating the use of artificial Cell Signaling Networks to implement computation, sign... more We are investigating the use of artificial Cell Signaling Networks to implement computation, signal processing and (or) control functionality. In the following sections we review a number of the research issues which this raises.
Artificial Life, 2008
This work investigates closure in Cell Signaling Networks, which is one research area within the ... more This work investigates closure in Cell Signaling Networks, which is one research area within the ESIGNET project 1 . We employ a string-based Artificial Chemistry based on Holland's broadcast language (Molecular Classifier System, Broadcast Language, or MCS.b). We present a series of experiments focusing on the emergence and evolution of self-maintaining molecular organizations. Such experiments naturally relate to similar studies conducted in artificial chemistries such as Tierra, Alchemy and Alpha-Universes. However, our results demonstrate some counter-intuitive outcomes, not indicated in previous literature. Each of these "unexpected" evolutionary dynamics (including an elongation catastrophe phenomenon) are examined and explained both informally and formally. We also demonstrate how the elongation catastrophe can be prevented using a multi-level selectional model of the MCS.b (which acts both at the molecular and cellular level). This work provides complementary insights into the understanding of evolutionary dynamics in minimal artificial chemistries.
Accommodating the complexity of parameter setting in multi-level evolutionary systems
Current approaches within the ESIGNET project (ESIGNET: Evolving Cell Signaling Networks in Silic... more Current approaches within the ESIGNET project (ESIGNET: Evolving Cell Signaling Networks in Silico), an EU FP6 project (contract number 12789 http://www.esignet.net) have used multilevel Evolutionary Algorithms (EA) to assist in the understanding of evolution and complexity of biochemical networks. These systems have utilized meta-level EAs to evolve parameters for base-level evolutionary searches. However, these EAs utilized a fixed pre-determined termination criterion for each of their searches. This constraint can potentially affect the computational cost and precision of the system. Our proposed solution uses a flexible approach to determining the termination criterion for the base level search and provides a fitness measure for the meta level search. We couple this termination criterion with a cost benefit function (the Quality Time Tradeoff function - QTT) which trades the solution quality against computation time. In this manner the system can achieve a level of accuracy to s...
China-Ireland International Conference on Information and Communications Technologies (CIICT 2008), 2008
This paper is concerned with the modeling and evolving of Cell Signaling Networks (CSNs) in silic... more This paper is concerned with the modeling and evolving of Cell Signaling Networks (CSNs) in silico. CSNs are complex biochemical networks responsible for the coordination of cellular activities. We examine the possibility to computationally evolve and simulate Artificial Cell Signaling Networks (ACSNs) by means of Evolutionary Computation techniques. From a practical point of view, realizing and evolving ACSNs may provide novel computational paradigms for a variety of application areas. For example, understanding some inherent properties of CSNs such as crosstalk may be of interest: A potential benefit of engineering crosstalking systems is that it allows the modification of a specific process according to the state of other processes in the system. This is clearly necessary in order to achieve complex control tasks. This work may also contribute to the biological understanding of the origins and evolution of real CSNs. An introduction to CSNs is first provided, in which we describe the potential applications of modeling and evolving these biochemical networks in silico. We then review the different classes of techniques to model CSNs, this is followed by a presentation of two alternative approaches employed to evolve CSNs within the ESIGNET project 1 . Results obtained with these methods are summarized and discussed.
Lecture Notes in Computer Science
The Broadcast Language is a programming formalism devised by Holland in 1975, which aims at impro... more The Broadcast Language is a programming formalism devised by Holland in 1975, which aims at improving the efficiency of Genetic Algorithms (GAs) during long-term evolution. The key mechanism of the Broadcast Language is to allow GAs to employ an adaptable problem representation. Fixed problem encoding is commonly used by GAs but may limit their performance in particular cases. This paper describes an implementation of the Broadcast Language and its application to modeling biochemical networks. Holland presented the Broadcast Language in his book "Adaptation in Natural and Artificial Systems" where only a description of the language was provided, without any implementation. Our primary motivation for this work was the fact that there is currently no published implementation of the Broadcast Language available. Secondly, no additional examination of the Broadcast Language and its applications can be found in the literature. Holland proposed that the Broadcast Language would be suitable for the modeling of biochemical models. However, he did not support this belief with any experimental work. In this paper, we propose an implementation of the Broadcast Language which is then applied to the modeling of a signal transduction network. We conclude the paper by proposing that with some refinements it will be possible to use the Broadcast Language to evolve biochemical networks in silico.
2009 IEEE Congress on Evolutionary Computation, 2009
We examine a potential role of signalling crosstalk in Artificial Cell Signalling Networks (ACSNs... more We examine a potential role of signalling crosstalk in Artificial Cell Signalling Networks (ACSNs). In this research, we regard these ACSNs or Artificial Biochemical Networks (ABNs) as collectively autocatalytic (i.e., closed) reaction networks being able to both self-maintain and to carry out a distinct signal processing function. These signalling crosstalk phenomena occur naturally when different biochemical networks become mixed together where a given molecular species may contribute simultaneously to multiple ACSNs. It has been reported in the biological literature, that crosstalk may have effects that are both constructive (e.g., coordinating cellular activities, multi-tasking) and destructive (e.g., premature programmed cell death). In this paper we demonstrate how crosstalk may enable distinct closed ACSNs to cooperate with other. From a theoretical point of view, this work may give new insights for the understanding of crosstalk in natural biochemical networks. From a practical point view, this investigation may provide novel applications of crosstalk in engineered ABNs.
Complex Adaptive Systems (CAS) are dynamical networks of interacting agents occurring in a variet... more Complex Adaptive Systems (CAS) are dynamical networks of interacting agents occurring in a variety of natural and artificial systems (eg cells, societies, stock markets). These complex systems have the ability to adapt, evolve and learn from experience. To study CAS, Holland proposed to employ agent-based systems in which Learning Classifier Systems (LCS) are used to determine the agents behavior and adaptivity. We argue that LCS are limited for the study of CAS: the rule-discovery mechanism is pre-specified and may limit the evolvability ...
All processes of life are dominated by networks of interacting biochemical components. The purpos... more All processes of life are dominated by networks of interacting biochemical components. The purpose of modeling these networks is manifold. From a theoretical point of view it allows the exploration of network structures and dynamics, to find emergent properties or to explain the organization and evolution of networks. From a practical point of view, in silico experiments can be performed that would be very expensive or impossible to achieve in the laboratory, such as hypothesis-testing with regard to knockout experiments or ...
Protocells are hypothesised as a transitional phase in the origin of life, prior to the evolution... more Protocells are hypothesised as a transitional phase in the origin of life, prior to the evolution of fully functional prokaryotic cells. The work reported here is being done in the context of the PACE project, which is investigating the fabrication of artificial protocells de novo. We consider here the important open question of whether or how articifial protocells (if or when they are successfully fabricated) might be applied as “computing” devices—what sort of computing might they be suitable for, and how might they be “programmed”? We also ...
This work investigates closure in Cell Signaling Networks, which is one research area within the ... more This work investigates closure in Cell Signaling Networks, which is one research area within the ESIGNET project. We employ a string-based Artificial Chemistry based on Holland's broadcast language (Molecular Classifier System, Broadcast Language, or MCS. b). We present a series of experiments focusing on the emergence and evolution of self-maintaining molecular organizations. Such experiments naturally relate to similar studies conducted in artificial chemistries such as Tierra, Alchemy and Alpha-Universes. However ...
2007 IEEE Congress on Evolutionary Computation, 2007
In this paper we present a novel cost benefit operator that assists multi level genetic algorithm... more In this paper we present a novel cost benefit operator that assists multi level genetic algorithm searches. Through the use of the cost benefit operator, it is possible to dynamically constrain the search of the base level genetic algorithm, to suit the user's requirements. Initially we review meta-evolutionary (multi-level genetic algorithm) approaches. We note that the current literature has abundant studies on meta-evolutionary GAs. However these approaches have not identified an efficient approach to termination of base GA search or a means to balance practical consideration such as quality of solution and the expense of computation. Our Quality time tradeoff operator (QTT) is user defined, and acts as a base level termination operator and also provides a fitness value for the meta-level GA. In this manner the amount of computation time spent on less encouraging configurations can be specified by the user. Our approach has been applied to a computationally intensive test problem which evaluates a large set of configuration settings for the base GAs. This approach should be applicable across a wide range of practical problems (e.g. routing, logistic and biomedical applications). number of different approaches have been taken to the development of Genetic Algorithms for different application areas. The genetic algorithms typically consist of a number of core parts: representation, fitness evaluation, crossover and mutation operators. Many differing forms of operators have been developed.
Complex Adaptive Systems (CAS) are dynamical networks of interacting agents which as a whole dete... more Complex Adaptive Systems (CAS) are dynamical networks of interacting agents which as a whole determine the behavior, adaptivity and cognitive ability of the system. CAS are ubiquitous and occur in a variety of natural and artificial systems (eg, cells, societies, stock markets). To study CAS, Holland proposed to employ an agent-based system in which Learning Classifier Systems (LCS) were used to determine the agents behavior and adaptivity. We argue that LCS are limited for the study of CAS: the rule-discovery ...
2006 1st Bio-Inspired Models of Network, Information and Computing Systems, 2006
Nature is a source of inspiration for computational techniques which have been successfully appli... more Nature is a source of inspiration for computational techniques which have been successfully applied to a wide variety of complex application domains. In keeping with this we examine Cell Signaling Networks (CSN) which are chemical networks responsible for coordinating cell activities within their environment. Through evolution they have become highly efficient for governing critical control processes such as immunological responses, cell cycle control or homeostasis. Realising (and evolving) Artificial Cell Signaling Networks (ACSNs) may provide new computational paradigms for a variety of application areas. Our abstraction of Cell Signaling Networks focuses on four characteristic properties distinguished as follows: Computation, Evolution, Crosstalk and Robustness. These properties are also desirable for potential applications in the control systems, computation and signal processing field. These characteristics are used as a guide for the development of an ACSN evolutionary simulation platform. In this paper we present a novel evolutionary approach named Molecular Classifier System (MCS) to simulate such ACSNs. The MCS that we have designed is derived from Holland's Learning Classifier System. The research we are currently involved in is part of the multi disciplinary European funded project, ESIGNET, with the central question of the study of the computational properties of CSNs by evolving them using methods from evolutionary computation, and to reapply this understanding in developing new ways to model and predict real CSNs.
Uploads
Papers by George Mitchell