Carlos M. Fernandes

Resilience to churn of a peer-to-peer evolutionary algorithm

In this paper we analyse the resilience of a peer-to-peer (P2P) evolutionary algorithm (EA) subject to the following dynamics: computing nodes acting as peers leave the system independently from each other causing a collective effect known as churn. Since the P2P EA has been designed to tackle large instances of computationally expensive problems, we will assess its behaviour under these conditions, by performing a scalability analysis in five different scenarios using the massively multimodal deceptive problem as a benchmark. In all cases, the P2P EA reaches the success criterion without a penalty on the runtime. We show that the key to the algorithm resilience is to ensure enough peers at the beginning of the experiment; even if some of them leave, those that remain contain enough information to guarantee a reliable convergence.

Varying the population size of artificial foraging swarms on time varying landscapes

In this paper we present a Swarm Search Algorithm with varying population of agents based on a previous model with fixed population which proved its effectiveness on several computation problems [6,7,8]. We will show that the variation of the population size provides the swarm with mechanisms that improves its self-adaptability and causes the emergence of a more robust self-organized behavior, resulting in a higher efficiency on searching peaks and valleys over dynamic search landscapes represented here by several three-dimensional mathematical functions that suddenly change over time.

Self-regulated population size in evolutionary algorithms

In this paper we analyze a new method for an adaptive variation of Evolutionary Algorithms (EAs) population size: the Self-Regulated Population size EA (SRP-EA). An empirical evaluation of the method is provided by comparing the new proposal with the CHC algorithm and other well known EAs with varying population. A fitness landscape generator was chosen to test and compare the algorithms: the Spears multimodal function generator. The performance of the algorithms was measured in terms of success rate, quality of the solutions and evaluations needed to attain them over a wide range of problem instances. We will show that SRP-EA performs well on these tests and appears to overcome some recurrent drawbacks of traditional EAs which lead them to local optima premature convergence. Also, unlike other methods, SRP-EA seems to self-regulate its population size according to the state of the search.

Effect of Noisy Fitness in Real-Time Strategy Games Player Behaviour Optimisation Using Evolutionary Algorithms

This paper investigates the performance and the results of an evolutionary algorithm (EA) specifically designed for evolving the decision engine of a program (which, in this context, is called bot) that plays Planet Wars. This game, which was chosen for the Google Artificial Intelligence Challenge in 2010, requires the bot to deal with multiple target planets, while achieving a certain degree of adaptability in order to defeat different opponents in different scenarios. The decision engine of the bot is initially based on a set of rules that have been defined after an empirical study, and a genetic algorithm (GA) is used for tuning the set of constants, weights and probabilities that those rules include, and therefore, the general behaviour of the bot. Then, the bot is supplied with the evolved decision engine and the results obtained when competing with other bots (a bot offered by Google as a sparring partner, and a scripted bot with a pre-established behaviour) are thoroughly analysed. The evaluation of the candidate solutions is based on the result of non-deterministic battles (and environmental interactions) against other bots, whose outcome depends on random draws as well as on the opponents’ actions. Therefore, the proposed GA is dealing with a noisy fitness function. After analysing the effects of the noisy fitness, we conclude that tackling randomness via repeated combats and reevaluations reduces this effect and makes the GA a highly valuable approach for solving this problem.

Optimizing player behavior in a real-time strategy game using evolutionary algorithms

This paper describes an Evolutionary Algorithm for evolving the decision engine of a bot designed to play the Planet Wars game. This game, which has been chosen for the Google Artificial Intelligence Challenge in 2010, requires that the artificial player is able to deal with multiple objectives, while achieving a certain degree of adaptability in order to defeat different opponents in different scenarios. The decision engine of the bot is based on a set of rules that have been defined after an empirical study. Then, an Evolutionary Algorithm is used for tuning the set of constants, weights and probabilities that define the rules, and, therefore, the global behavior of the bot. The paper describes the Evolutionary Algorithm and the results attained by the decision engine when competing with other bots. The proposed bot defeated a baseline bot in most of the playing environments and obtained a ranking position in top-20% of the Google Artificial Intelligence competition.

niGAVaPS — outbreeding in genetic algorithms

This paper presents niGAVaPS (non-incest Genetic Algorithm with Variable Population Size), a genetic algorithm that mimics some mechanisms of evolution in natural environment populations. It prevents incest by forbidding the recombination of closely related individuals, based on their ancestry. The degree of ancestry considered is adjustable, niGAVaPS tunes the population size parameter according to the state of the search process. Each chromosome has a lifetime, which corresponds to the number of generations in which it will remain in the population. This value is calculated according to the chromosome fitness and population characteristics at the moment of its creation. Preventing incest supposedly helps maintaining population diversity thus avoiding premature convergence. The purpose of our work is to test these ideas on a range of problems, as far as algorithm convergence and quality of solutions found are concerned.

High school weekly timetabling by evolutionary algorithms

This paper describes a method for generating high school timetables using an Evolutionary Algorithm (E.A.). Given c classes. t teachers and r classrooms it is required to build a set of c+t+r timetables based on the needs of the school and satisfying constraints in the assignment of the lessons. A problem specific chromosome representation and the use of a repair function during titness evaluation helps the algorithm, keeping the search close to valid solutions. Our new operator, ‘Bad Genes Mutation’, greatly improved the algorithm’s speed and results. Test results on a large high school are presented. 1 THE SCHOOL TlMETABLlNG PROBLEM In this section we describe in detail the schooltimetabling problem and discuss the hard and soft [Z] constraints considered in the assignment of a lesson. 1.7 Defining the Problem Conditions We define time slot duration as the greatest common factor of all the different lesson durations i.e. if we only have sixty-minute and ninety-minute lessons then the time slot duration will be th’irty minutes. We define the dimension (number of timeslots) of each timetable as: dimension = rrulnber-of-days-wit/l-lessons * number-hne_slotsJerJay The most common division of timetables into time slots used in schools is a sixty-minute time slot. ten times a day and five days a week, as shown in Figure 1. p-ission ~0 m&e di&al or hard copies Of dt Oc @ Oftis wk for p-d 0~ classroom use is pnttd without fee pm”ia thJ copime no( made OT distributed fOr FOffi 01 COll’UnS~

Using assortative mating in genetic algorithms for vector quantization problems

&an(a8e and that copies bear this notice and the full citation on he first page. To copy otherwise, to republish to P.04 on Seem 0~ to redi&bute to lists, requires prior Specific pennlmm a&of a fee. SAC 99, San Antonio, Texas 61998 ACM l-58113-0864991ooO1 S5.00 Fernando Melicio’ Agostinho Rosa I.S.E.L. LaSEEB-ISR-IST R. Conselheiro Emidio Navarro Av. Rovisco Pais, I. TN 6.21 1900 Lisboa I O49IO0 Lisboa Codex Portugal Portugal Fmelicio@isel.ot Acrosa@isr.ist.utl.ot Fig I Map for the assipnent t,f the lesssscms. Each time SIOI is asswiuted with a number

Improving genetic algorithms performance via deterministic population shrinkage

In nature, some species mate according to their phenotype similarity. The Assortative Mating Genetic Algorithm (AMGA) mimics some mechanisms of reproduction in natural environments. The main difference between AMGA and the Standard GA (SGA) is the selection of the parents in the crossover operators. We develop a similarity measure for the Vector Quantization problem and we show that the application of AMGA to some instances of this problem reduces the number of times that the algorithm becomes trapped in local optima. We also present results that show that AMGA keeps a higher level of genetic diversity than the SGA.

Self-adjusting the intensity of assortative mating in genetic algorithms

Despite the intuition that the same population size is not needed throughout the run of an Evolutionary Algorithm (EA), most EAs use a fixed population size. This paper presents an empirical study on the possible benefits of a Simple Variable Population Sizing (SVPS) scheme on the performance of Genetic Algorithms (GAs). It consists in decreasing the population for a GA run following a predetermined schedule, configured by a speed and a severity parameter. The method uses as initial population size an estimation of the minimum size needed to supply enough building blocks, using a fixed-size selectorecombinative GA converging within some confidence interval toward good solutions for a particular problem. Following this methodology, a scalability analysis is conducted on deceptive, quasi-deceptive, and non-deceptive trap functions in order to assess whether SVPS-GA improves performances compared to a fixed-size GA under different problem instances and difficulty levels. Results show several combinations of speed-severity where SVPS-GA preserves the solution quality while improving performances, by reducing the number of evaluations needed for success.