Figure 5 - uploaded by Giovanni Squillero
Content may be subject to copyright.
Source publication
In this paper, Core War, a very peculiar game popular in mid 80's, is exploited as a benchmark to improve the μGP, an evolutionary algorithm able to generate touring-complete, realistic assembly programs. Two techniques were analyzed: coevolution and a modified island model. Experimental results showed that the former is essential in the beginning...
Context in source publication
Similar publications
Rapid intensification in tropical cyclones occur where there is dramatic change in wind-intensity over a short period of time. In this paper , recurrent neural networks trained with cooperative neuro-evolution method are used for prediction of rapid intensification in tropical cyclones in the South Pacific region. An analysis of the tropical cyclon...
Time series prediction relies on past data points to make robust predictions. The span of past data points is important for some applications since prediction will not be possible unless the minimal timespan of the data points is available. This is a problem for cyclone wind-intensity prediction, where prediction needs to be made as a cyclone is id...
We investigate Bak-Sneppen coevolution models on scale-free networks with various degree exponents gamma including random networks. For gamma>3 , the critical fitness value f(c) approaches a nonzero finite value in the limit N --> infinity, whereas f(c) approaches zero as 2<gamma< or =3. These results are explained by showing analytically f(c) (N)...
Cooperative coevolution decomposes a large problem into its subcomponents and uses evolutionary algorithms for solving them in order to gradually solve the large problem. This paper uses cooperative coevolution framework for training recurrent neural networks for grammatical inference problems. In the past, different encoding schemes were used to b...
Citations
... Exploration and generation of CoreWar warriors, assisted by computers, has become increasingly popular in the recent years. Majority of work, however, has concentrated on warrior parameter optimization [7] and the evolution of competitive warriors [8,5,9]. Exploratory analysis (albeit motivated by warrior evolution), by means of automatic categorization based on the analysis of execution frequencies of certain instruction types during simulation, was performed, with some results available in [2], but with no published findings. ...
... Starting from perhaps the earliest efforts of Samuel (1959) with checkers (also known as draughts), research has been conducted in various forms of computational intelligence (e.g., neural networks and evolutionary computation) for a wide variety of n-player zero-sum and nonzero-sum games. A complete review of these efforts is beyond the scope of this paper; however the breadth of study can be appreciated by reviewing the numerous contributions to games research using the iterated prisoner's dilemma (e.g., Axelrod, 1987;Fogel, 1993;Fogel, 1995;Harrald and Fogel, 1996;Darwen and Yao, 2000;Chong and Yao, 2004;Chong and Yao, 2005;Ishibuchi and Namikawa, 2005;Franken and Engelbrecht, 2005 and others), general game theory and evolutionary stable strategies ( Fogel et al., 1997;Fogel et al., 1998;Fogel and Beyer, 2000;Ficici et al., 2005 and others), board games such as checkers Fogel, 1999a,b, 2001;Fogel, 2002;Hughes, 2003;Franken and Engelbrecht, 2003;Cho, 2003, 2005;Hughes, 2005), chess ( Kendall and Whitwell, 2001;Fogel et al., 2004a,b), Othello (Moriarty and Miikkulainen, 1995), backgammon (Pollack and Blair, 1998;Darwen, 2001), versions of Go ( Richards et al., 1998Richards et al., , 2001Kendall et al., 2004;Runnarson and Lucas, 2005;Lubberts and Miikkulainen, 2001), RISK ( Vaccaro and Guest, 2005), Monopoly (Frayn, 2005), and other games such as core wars ( Corno et al., 2003Corno et al., , 2004Corno et al., 2005a,b), card games ( Kendall and Smith, 2003;Fogel, 2004), combat and other video games ( Gallagher and Ryan, 2003;Stanley et al., 2005;Louis and Miles, 2005;Hong and Cho, 2005;Tanev et al., 2005;Togelius and Lucas, 2005;Parker et al., 2005), and many others (see also Kendall and Lucas, 2005;Fogel et al., 2005). ...
Entertainment software developers face significant challenges in designing games with broad appeal. One of the challenges concerns creating nonplayer (computer-controlled) characters that can adapt their behavior in light of the current and prospective situation, possibly emulating human behaviors. This adaptation should be inherently novel, unrepeatable, yet within the bounds of realism. Evolutionary algorithms provide a suitable method for generating such behaviors. This paper provides background on the entertainment software industry, and details a prior and current effort to create a platform for evolving nonplayer characters with genetic and behavioral traits within a World War I combat flight simulator.
... In 2003, Corno, Sanchez and Squillero used corewar as a test bench to enhance the µGP, a generic assembly-level program generator [4]. The game was exploited to evaluate the effectiveness of new evolutionary methodologies: the results attained by evolved warriors were used as a feedback for the adopted selection schemes and operators. ...
... This paper is an enhancement of [4] and it shares the same goal: exploit a game to investigate new evolutionary techniques. It presents a new migration model that exploits the polarization effect and a new hierarchical coarse-grained approach applicable whenever the final goal can be seen as a combination of semi-independent sub goals. ...
... According to the suggestion of [5], the instruction library for the corewar has been rewritten from scratch. While in [4] a single macro may encode several different redcode instructions as in Figure 2, in the instruction library adopted each macro corresponds to one single redcode instruction. ...
This paper analyzes corewar, a very peculiar computer game popular in mid 80's where different programs fight in the memory of a virtual computer. The μGP, an evolutionary assembly-program generator, is used to evolve efficient programs, and the game is exploited to evaluate new evolutionary techniques. The paper introduces a new migration model that exploits the polarization effect and a new hierarchical coarse-grained approach applicable whenever the final goal can be seen as a combination of semi-independent sub goals. Additionally, two very general enhancements are proposed. Analyzed techniques are orthogonal and broadly applicable to different real-life contexts. Experimental results show that all these techniques are able to outperform a previous approach.
... Evolved CoreWar programs are still superior to humancoded programs in some of the basic confrontation environments. However, these evolvers have failed to produce programs of higher complexity and there is usually little diversity in the evolved sets [3]. Program optimization does not have to be related to changes in the code structure. ...
In this paper, an optimizer for programs written in an assembly-like language called Redcode is presented. Relevance of code optimization in evolutionary program creation strategies and code categorization is discussed. CoreWar Optimizer is the first user-friendly optimization tool for CoreWar programs offering various optimization methods and a carefully picked benchmark. The methods at the user's disposal are: random, modified hill climbing algorithm, simulated annealing, predator-prey particle swarm optimization and genetic algorithms. All these methods use a speed-up trick which drives the value optimization across three fitness landscapes instead of just one.
... The genetic programming literature consistently cites the importance of maintaining diversity as being crucial in avoiding convergence toward local optima and there are several different possible strategies to promote diversities, including non-standard selection, mating, or a replacement strategy. Indeed, µGP performances were already enhanced exploiting coarse-grained approaches, and such geographical distributions of individual are known to promote diversity [30] [31]. ...
The development of computer-generated ecosystem simulations are becoming more common due to the greater computational capabilities of machines. Because natural ecosystems are very complex, simplifications are required for implementation. This simulation environment offer a global view of the system and generate a lot of data to process and analyse, which are difficult or impossible to do in nature. 3D simulations, besides of the scientific advantages in experiments, can be used for presentation , educational and entertainment purposes too. In our simulated framework (Animal Farm) we have implemented a few basic animal behaviors and mechanics to observe in 3D. All animals are controlled by an individual logic model, which determines the next action of the animal, based on their needs and surrounding environment.
Core War is a game where two or more programs, called warriors, are executed in the same memory area by a time-sharing processor. The final goal of each warrior is to crash the others by overwriting them with illegal instructions. The game was popularized by A. K. Dewdney in his Scientific American column in the mid-1980s. In order to automatically devise strong warriors, μGP, a test program generation algorithm, was extended with the ability to assimilate existing code and to detect clones; furthermore, a new selection mechanism for promoting diversity independent from fitness calculations was added. The evolved warriors are the first machine-written programs ever able to become King of the Hill (champion) in all four main international Tiny Hills. This paper shows how playing Core War may help generate effective test programs for validation and test of microprocessors. Tackling a more mundane problem, the described techniques are currently being exploited for the automatic completion and refinement of existing test programs. Preliminary experimental results are reported.
