Results 1  10
of
15
Evolutionary neurocontrollers for autonomous mobile robots
 NEURAL NETWORKS
, 1998
"... In this article we describe a methodology for evolving neurocontrollers of autonomous mobile robots without human intervention. The presentation, which spans from technological and methodological issues to several experimental results on evolution of physical mobile robots, covers both previous and ..."
Abstract

Cited by 98 (10 self)
 Add to MetaCart
(Show Context)
In this article we describe a methodology for evolving neurocontrollers of autonomous mobile robots without human intervention. The presentation, which spans from technological and methodological issues to several experimental results on evolution of physical mobile robots, covers both previous and recent work in the attempt to provide a uni ed picture within which the reader can compare the effects of systematic variations on the experimental settings. After describing some key principles for building mobile robots and tools suitable for experiments in adaptive robotics, we give an overview of different approaches to evolutionary robotics and present our methodology. We start reviewing two basic experiments showing that different environments can shape very different behaviors and neural mechanisms under very similar selection criteria. We then address the issue of incremental evolution in two different experiments from the perspective of changing environments and robot morphologies. Finally, we investigate the possibility of evolving plastic neurocontrollers and analyze an evolved neurocontroller that relies on fast and continuously changes synapses characterized by dynamic stability. We conclude by reviewing the implications of this methodology for engineering, biology, cognitive science, and artificial life, and point at future directions of research.
LibGA: A userfriendly workbench for orderbased genetic algorithm research
 Proceedings of the 1993 ACM/SIGAPP Symposium on Applied Computing
, 1993
"... Over the years there has been several packages developed that provide a workbench for genetic algorithm (GA) research. Most of these packages use the generational model inspired by GENESIS. A few have adopted the steadystate model used in Genitor. Unfortunately, they have some de ciencies when work ..."
Abstract

Cited by 37 (17 self)
 Add to MetaCart
(Show Context)
Over the years there has been several packages developed that provide a workbench for genetic algorithm (GA) research. Most of these packages use the generational model inspired by GENESIS. A few have adopted the steadystate model used in Genitor. Unfortunately, they have some de ciencies when working with orderbased problems such aspacking, routing, and scheduling. This paper describes LibGA, which was developed speci cally for orderbased problems, but which also works easily with other kinds of problems. It offers an easy to use `userfriendly ' interface and allows comparisons to be made between both generational and steadystate genetic algorithms for a particular problem. It includes a variety of genetic operators for reproduction, crossover, and mutation. LibGA makes it easy to use these operators in new ways for particular applications or to develop and include new operators. Finally, it o ers the unique new feature of a dynamic generation gap.
Convergence of non{elitist strategies
 In Proceedings of the First IEEE Conference on Computational Intelligence
, 1994
"... Abstract  This paper o ers su cient conditions to prove global convergence of non{elitist evolutionary algorithms. If these conditions can be applied they yield bounds of the convergence rate as a by{product. This is demonstrated by an example that can be calculated exactly. ..."
Abstract

Cited by 32 (4 self)
 Add to MetaCart
(Show Context)
Abstract  This paper o ers su cient conditions to prove global convergence of non{elitist evolutionary algorithms. If these conditions can be applied they yield bounds of the convergence rate as a by{product. This is demonstrated by an example that can be calculated exactly.
The SAGA Cross: The Mechanics of Recombination for Species with Variablelength Genotypes
, 1992
"... ..."
Parallel Approaches to Stochastic Global Optimization
 In Parallel Computing: From Theory to Sound Practice, W. Joosen and E. Milgrom, Eds., IOS
, 1992
"... In this paper we review parallel implementations of some stochastic global optimization methods on MIMD computers. Moreover, we present a new parallel version of an Evolutionary Algorithm for global optimization, where the inherent parallelism can be scaled to obtain a reasonable processor utilizati ..."
Abstract

Cited by 14 (5 self)
 Add to MetaCart
(Show Context)
In this paper we review parallel implementations of some stochastic global optimization methods on MIMD computers. Moreover, we present a new parallel version of an Evolutionary Algorithm for global optimization, where the inherent parallelism can be scaled to obtain a reasonable processor utilization. For this algorithm the convergence to the global optimum with probability one can be assured. Test results concerning speed up and reliability are given. 1 Introduction Many real world problems in engineering and economics can be formulated as optimization problems, in which the objective function is multimodal, i.e. the problem possesses many local minima. Compared to the number of methods designed to determine a local minimum, there are only a few methods which attempt to find the global minimum (see [52] for a survey). Although there are some special cases where the global optimum can be found (see [26]) the general case is unsolvable. This paper will be restricted to the more gener...
Optimal Surface Reconstruction from Digitized Point Data using CI Methods
 UNIVERSITY OF DORTMUND
, 1997
"... In many scientific and technological endeavors, a threedimensional solid must be reconstructed from digitized point data. This paper presents three solutions to the problem of reconstructing smooth surfaces using triangular tiles. The presented algorithms differ in their strategic approach. Here, t ..."
Abstract

Cited by 4 (2 self)
 Add to MetaCart
In many scientific and technological endeavors, a threedimensional solid must be reconstructed from digitized point data. This paper presents three solutions to the problem of reconstructing smooth surfaces using triangular tiles. The presented algorithms differ in their strategic approach. Here, two deterministic algorithms and one nondeterministic CI (computational intelligence) strategy will be described. In order to compare triangulations, two quality criteria will be introduced.
A Heuristic for Improved Genetic Bin Packing
 INFORMATION PROCESSING LETTERS
, 1993
"... The bin packing optimization problem packs a set of objects into a set of bins so that the amount of wasted space is minimized. The bin packing problem has many important applications. These include multiprocessor scheduling, resource allocation, and realworld planning, packing, routing, and sche ..."
Abstract

Cited by 4 (2 self)
 Add to MetaCart
(Show Context)
The bin packing optimization problem packs a set of objects into a set of bins so that the amount of wasted space is minimized. The bin packing problem has many important applications. These include multiprocessor scheduling, resource allocation, and realworld planning, packing, routing, and scheduling optimization problems. The bin packing problem is NPcomplete [4]. Since there is therefore little hope in finding an efficient deterministic solution to the bin packing problem, approximation methods have been developed. The advantage of these methods is that they have guaranteed packing performance bounds. A survey of approximation algorithms for bin packing and their respective performance bounds are reported by Garey and Johnson [5] in the one dimensional case, Coffman et al. [1] in the two dimensional case, and Li and Cheng [8] in the three dimensional case. In many practical applications of bin packing, a small improvement in packing efficienc
Solving NPComplete Problems in RealTime System Design by Multichromosome Genetic Algorithms
 In Proceedings of the SIGPLAN 1997 Workshop on Languages, Compilers, and Tools for RealTime Systems
, 1997
"... Most problems in the design of realtime applications like task allocation or scheduling belong to the class of NPcomplete problems and can be solved efficiently only by heuristics. ..."
Abstract

Cited by 4 (1 self)
 Add to MetaCart
(Show Context)
Most problems in the design of realtime applications like task allocation or scheduling belong to the class of NPcomplete problems and can be solved efficiently only by heuristics.
Honors
"... We use the Genetic Algorithm (GA), a heuristic search and optimization technique inspired by biological evolution, to search for or \evolve " models of partially known nonlinear dynamical systems. We use certain assumptions about the class of \goal " systems (those being modeled), to build ..."
Abstract

Cited by 3 (0 self)
 Add to MetaCart
(Show Context)
We use the Genetic Algorithm (GA), a heuristic search and optimization technique inspired by biological evolution, to search for or \evolve " models of partially known nonlinear dynamical systems. We use certain assumptions about the class of \goal " systems (those being modeled), to build constraints into our \model " systems, which consist of functions represented by tables of numbers. Further knowledge is incorporated into our error metric, which is de ned (only) for autonomous dynamical systems. Because we assume that both model and goal systems are autonomous (invariant with respect to translation in time), it is possible to compare them on the basis of the geometry of their respective phase portraits. Thus we formulate a measure, based on phase portrait geometry, of the error or \distance " between dynamical systems. By minimizing the distance separating the model from the goal system, the Genetic Algorithm is usually able to nd an approximation of the goal system. We haveused GAdget, our objectoriented implementation of the GA, to
Chromosome reduction in Genetic algorithms
, 1994
"... Genetic Algorithms [8, 9]havebeen very successful as adaptive search techniques on a wide variety of problems. However, the GA has di culty with some problems, such as those with rugged tness landscapes or those which require nonbinary genome representations. Recent research has focused on new theo ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
Genetic Algorithms [8, 9]havebeen very successful as adaptive search techniques on a wide variety of problems. However, the GA has di culty with some problems, such as those with rugged tness landscapes or those which require nonbinary genome representations. Recent research has focused on new theories, models and techniques for improving the performance of GAs on di cult problems. Many new ideas have