Abstract not found

Tag is a children's game based on symmetrical pursuit and evasion. In the experiments described here, control programs for mobile agents (simulated vehicles) are evolved based on their skill at the game of tag. A player's fitness is determined by how well it performs when placed in competition with several opponents chosen randomly from the coevolving population of players. In the beginning, the quality of play is very poor. Then slightly better strategies begin to exploit the weaknesses of others. Through evolution, guided by competitive fitness, increasingly better strategies emerge over time. 1. Introduction Many of us remember playing the game of tag as children. Tag is played by two or more, one of whom is designated as it. The it player chases the others, who all try to escape. Tag is a simple contest of pursuit and evasion. These activities are common in the natural world, most predatorprey interactions involve pursuit and evasion. Tag also includes an aspect of role-reversal, b...

A key concern in genetic programming (GP) is the size of the state--space which must be searched for large and complex problem domains. One method to reduce the state--space size is by using Strongly Typed Genetic Programming (STGP). We applied both GP and STGP to construct cooperation strategies to be used by multiple predator agents to pursue and capture a prey agent on a grid--world. This domain has been extensively studied in Distributed Artificial Intelligence (DAI) as an easy--to--describe but difficult--to--solve cooperation problem. The evolved programs from our systems are competitive with manually derived greedy algorithms. In particular the STGP paradigm evolved strategies in which the predators were able to achieve their goal without explicitly sensing the location of other predators or communicating with other predators. This is an improvement over previous research in this area. The results of our experiments indicate that STGP is able to evolve programs that perform sign...

Abstract not found

The structure of the fitness landscape on which genetic programming operates is examined. The landscapes of a range of problems of known difficulty are analyzed in an attempt to determine which landscape measures correlate with the difficulty of the problem. The autocorrelation of the fitness values of random walks, a measure which has been shown to be related to perceived difficulty using other techniques, is only a weak indicator of the difficulty as perceived by genetic programming. All of these problems show unusually low autocorrelation. Comparison of the range of landscape basin depths at the end of adaptive walks on the landscapes shows good correlation with problem difficulty, over the entire range of problems examined. I. INTRODUCTION Genetic Programming, like all algorithms which depend on some form of evolutionary adaptation, operates within the context of a fitness landscape. The concept of a fitness landscape, introduced by Sewell Wright [1932], refers to the mapping from ...

The Push programming language was developed for use in genetic and evolutionary computation systems, as the representation within which evolving programs are expressed. It has been used in the production of several significant results, including results that were awarded a gold medal in the Human Competitive Results competition at GECCO-2004. One of Push’s attractive features in this context is its transparent support for the expression and evolution of modular architectures and complex control structures, achieved through explicit code self-manipulation. The latest version of Push, Push3, enhances this feature by permitting explicit manipulation of an execution stack that contains the expressions that are queued for execution in the interpreter. This paper provides a brief introduction to Push and to execution stack manipulation in Push3. It then presents a series of examples in which Push3 was used with a simple genetic programming system (PushGP) to evolve programs with non-trivial control structures.

The goal of automatic programming is to create, in an automated way, a computer program that enables a computer to solve a problem. Ideally, an automatic programming system should require that the user pre-specify as little as possible about the problem. In particular, it is desirable that the user not be required to specify the size and shape (i.e., the architecture) of the ultimate solution to the problem before applying the technique. This paper describes how the biological theory of gene duplication described in Susumu Ohno's provocative book, Evolution by Means of Gene Duplication, was brought to bear on a vexatious problem from the domain of automated machine learning in the computer science field. The resulting biologically-motivated approach using six new architecture-altering operations enables genetic programming to automatically discover the size and shape of the solution at the same time as it is evolving a solution to the problem. Genetic programming with the architecture-altering operations was used to evolve a computer program to classify a given protein segment as being a transmembrane domain or non-transmembrane area of the protein (without biochemical knowledge, such as

The design and development ofbehavioral strategies to coordinate the actions of multiple agents is a central issue in multiagent systems research. We propose a novel approach ofevolving, rather than handcrafting, behavioral strategies. The evolution scheme usedisavariant ofthe Genetic Programming (GP) paradigm. As a proof of principle, we evolve behavioral strategies in the predator{prey domain that has been studied widely in the DistributedArti cial Intelligence community. Weusethe GPto evolve behavioral strategies for individual agents, as prior literature claims that communication between predators is not necessary for successfully capturing the prey. Theevolved strategy, when used by each predator, performs better than all but oneofthe handcrafted strategies mentioned in literature. We analyze the shortcomings of each ofthese strategies. The next set of experiments involve co{evolving predators and prey. Toour surprise, a simple prey strategy evolves that consistently evades all of the predator strategies. We analyze the implications of the relative successes of evolution in the two sets of experiments and comment onthe nature of domains for which GPbasedevolutionisaviable mechanism for generating coordination strategies. We conclude withourdesign for concurrent evolution of multiple agent strategies in domains where agents need to communicate with eachother to successfully solve a common problem.

In general, the machine learning process can be accelerated through the use of additional knowledge about the problem solution. For example, monomorphic typed Genetic Programming (GP) uses type information to reduce the search space and improve performance. Unfortunately, monomorphic typed GP also loses the generality of untyped GP: the generated programs are only suitable for inputs with the specified type. Polymorphic typed GP improves over monomorphic and untyped GP by allowing the type information to be

Koza has previously shown that the power of a genetic programming system can often be enhanced by allowing for the simultaneous evolution of a main program and a collection of automatically defined functions (ADFs). In this paper I show how related techniques can be used to simultaneously evolve a collection of automatically defined macros (ADMs). I show how ADMs can be used to produce new control structures during the evolution of a program, and I present data showing that ADMs sometimes provide a greater benefit than do ADFs. I discuss the characteristics of problems that may benefit most from the use of ADMs, or from architectures that include both ADFs and ADMs, and I discuss directions for further research. Introduction Modern programming languages support the production of structured, modular programs through several mechanisms including subroutines, coroutines, and macros. Koza has shown that the power of a genetic programming system can often be enhanced by allowing for the ...