Evolving a neurocontroller through a process of embryogeny
نویسنده
چکیده
We introduce a model of cellular growth that generates neurocontrollers capable of guiding simple simulated agents in a harvesting task. The morphogenesis of the neurocontroller is itself controlled by an evolved artificial neural network. The neural network operates only on local variables and chemical concentrations and is thought as a flexible model of a gene regulatory system and cell metabolism. The model is designed in order to increase the evolvability of the growth mechanism, which constitutes a serious issue in artificial embryogeny. Also, to increase the flexibility of development, organisms are grown in embryonal stages, which allow an incremental refinement of development. Neurocontrollers are organized in horizontal layers, with vertical input and output pathways. Within the same layer, neurons can have only local connections. On one side this limits the information needed for routing and on the other makes the system easy to implement in hardware. Results show that the system is capable of developing appropriate neurocontrollers in most of the evolutionary runs.
منابع مشابه
Evolution of Neural Network’s Architecture through Symbiotic Neuroevolution
In this paper an extension of SANE that simultaneously evolves the weights and architecture of an MLP neural network is presented. The symbiotic adaptive neuroevolution (SANE) system coevolves a population of neurons that cooperate to form a functioning neural network. Evolutionary Strategies (ES) is applied to evolve the network weights. In order to increase the evolving system performance and...
متن کاملThe ABCs of Evolutionary Design: Investigating the Evolvability of Embryogenies for Morphogenesis
Previous work investigated three types of embryogeny: External, Explicit, and Implicit. This paper focuses on the two most interesting embryogenies: explicit and implicit, investigating the evolvability and scalability of both embryogenies for morphogenesis. The problem set is that of evolving certain predefined shapes letters of the alphabet. The results show that both embryogenies are good at...
متن کاملVector Field Embryogeny
We present a novel approach toward evolving artificial embryogenies, which omits the graph representation of gene regulatory networks and directly shapes the dynamics of a system, i.e., its phase space. We show the feasibility of the approach by evolving cellular differentiation, a basic feature of both biological and artificial development. We demonstrate how a spatial hierarchy formulation ca...
متن کاملEvolving ANNs for Spacecraft Rendezvous and Docking
We here approach the problem of designing a controller for automatic rendezvous and docking (AR&D). As a first step towards a fully reactive neurocontroller (based on state feedback), an artificial neural network is trained offline by a fitness based genetic algorithm to fulfill the docking task. Its performance using one specific docking case, with fixed initial and boundary conditions, is com...
متن کاملEvolving Small Neurocontrollers with Self-Organized Compact Encoding
This article presents a novel method for the evolution of artificial autonomous agents with small neurocontrollers. It is based on adaptive, self-organized compact genotypic encoding (SOCE) generating the phenotypic synaptic weights of the agent's neurocontroller. SOCE implements a parallel evolutionary search for neurocontroller solutions in a dynamically varying and reduced subspace of the or...
متن کامل