Juan Luis Jiménez Laredo

EvAg: a scalable peer-to-peer evolutionary algorithm

This paper studies the scalability of an Evolutionary Algorithm (EA) whose population is structured by means of a gossiping protocol and where the evolutionary operators act exclusively within the local neighborhoods. This makes the algorithm inherently suited for parallel execution in a peer-to-peer fashion which, in turn, offers great advantages when dealing with computationally expensive problems because distributed execution implies massive scalability. In this paper we show another advantage of this algorithm: We experimentally demonstrate that it scales up better than traditional alternatives even when executed in a sequential fashion. In particular, we analyze the behavior of several EAs on well-known deceptive trap functions with varying sizes and levels of deceptiveness. The results show that the new EA requires smaller optimal population sizes and fewer fitness evaluations to reach solutions. The relative advantage of the new EA is more outstanding as problem hardness and size increase. In some cases the new algorithm reduces the computational efforts of the traditional EAs by several orders of magnitude.

Asynchronous distributed genetic algorithms with Javascript and JSON

In a connected world, spare CPU cycles are up for grabs, if you only make its obtention easy enough. In this paper we present a distributed evolutionary computation system that uses the computational capabilities of the ubiquituous Web browser. Asynchronous Javascript and JSON (Javascript object notation, a serialization protocol) allows anybody with a Web browser (that is, mostly everybody connected to the Internet) to participate in a genetic algorithm experiment with little effort, or none at all. Since, in this case, computing becomes a social activity and is inherently impredictable, in this paper we will explore the performance of this kind of virtual computer by solving simple problems such as the royal road function and analyzing how many machines and evaluations it yields. We will also examine possible performance bottlenecks and how to solve them, and, finally, issue some advice on how to set up this kind of experiments to maximize turnout and, thus, performance. The experiments show that we we can obtain high, and to a certain point, reliable performance from volunteer computing based on AJAJ, with speedups of up to several (averaged) machines.

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.

Browser-based distributed evolutionary computation: performance and scaling behavior

The challenge of ad-hoc computing is to find the way of taking advantage of spare cycles in an efficient way that takes into account all capabilities of the devices and inter connections available to them. In this paper we explore distributed evolutionary computation based on the Ruby on Rails framework, which overlays a Model-View-Controller on evolutionary computation. It allows anybody with a web browser (that is, mostly everybody connected to the Internet) to participate in an evolutionary computation experiment. Using a straight forward farming model, we consider different factors, such as the size of the population used. We are mostly interested in how they impact on performance, but also the scaling behavior when a non-trivial number of computers is applied to the problem. Experiments show the impact of different packet sizes on performance, as well as a quite limited scaling behavior, due tothe characteristics of the server. Several solutions for that problem are proposed.

Implementation matters: programming best practices for evolutionary algorithms

While a lot of attention is usually devoted to the study of different components of evolutionary algorithms or the creation of heuristic operators, little effort is being directed at how these algorithms are actually implemented. However, the efficient implementation of any application is essential to obtain a good performance, to the point that performance improvements obtained by changes in implementation are usually much bigger than those obtained by algorithmic changes, and they also scale much better. In this paper we will present and apply usual methodologies for performance improvement to evolutionary algorithms, and show which implementation options yield the best results for a certain problem configuration and which ones scale better when features such as population or chromosome size increase.

Evolving bot AI in Unreal

This paper describes the design, implementation and results of an evolutionary bot inside the PC game UnrealTM, that is, an autonomous enemy which tries to beat the human player and/or some other bots. The default artificial intelligence (AI) of this bot has been improved using two different evolutionary methods: genetic algorithms (GAs) and genetic programming (GP). The first one has been applied for tuning the parameters of the hard-coded values inside the bot AI code. The second method has been used to change the default set of rules (or states) that defines its behaviour. Both techniques yield very good results, evolving bots which are capable to beat the default ones. The best results are yielded for the GA approach, since it just does a refinement following the default behaviour rules, while the GP method has to redefine the whole set of rules, so it is harder to get good results.

Cloud-based evolutionary algorithms: An algorithmic study

This paper presents a cloud-computing based evolutionary algorithm using a synchronous storage service as pool for exchange information among population of solutions. The multi-computer was composed of several normal PCs or laptops connected via Wifi or Ethernet. In this work the effect of how the distributed evolutionary algorithm reached the solution when new PCs was added was tested whether that effect also translates to the algorithmic performance of the algorithm. To this end different (and hard) problems was addressed using the proposed multi-computer, analyzing the effects that the automatic load-balancing and synchronization had on the speed of algorithm successful, and analyzing how the number of evaluation per second increases when the multi-computer includes new nodes. The measure used for the analysis was number of evaluation per second which was increased when the multi-computer includes new nodes. The algorithm solved the proposed problems and it was viable to run it in homogeneous or heterogeneous platforms. The experiments includes two problems and different configuration for the distributed evolutionary algorithm in order to check the results of the algorithm for several rates of information exchange with the selected storage service. Results shows that the system is viable with homogeneous or heterogeneous nodes and there is no significative differences for the synchronous storage services we have tested. But when the problem is harder, and the threads of the algorithm does not stop for each information exchange (migration of individual from one population to another one), the differences of using a specific service became significative in terms of success of the algorithm.

Improving genetic algorithms performance via deterministic population shrinkage

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.

Testing the Intermediate Disturbance Hypothesis: Effect of Asynchronous Population Incorporation on Multi-Deme Evolutionary Algorithms

In P2P and volunteer computing environments, resources are not always available from the beginning to the end, getting incorporated into the experiment at any moment. Determining the best way of using these resources so that the exploration/exploitation balance is kept and used to its best effect is an important issue. The Intermediate Disturbance Hypothesis states that a moderate population disturbance (in any sense that could affect the population fitness) results in the maximum ecological diversity. In the line of this hypothesis, we will test the effect of incorporation of a second population in a two-population experiment. Experiments performed on two combinatorial optimization problems, MMDP and P-Peaks , show that the highest algorithmic effect is produced if it is done in the middle of the evolution of the first population; starting them at the same time or towards the end yields no improvement or an increase in the number of evaluations needed to reach a solution. This effect is explained in the paper, and ascribed to the intermediate disturbanceproduced by first-population immigrants in the second population.

P2P Evolutionary Algorithms: A Suitable Approach for Tackling Large Instances in Hard Optimization Problems

In this paper we present a distributed Evolutionary Algorithm (EA) whose population is structured using newscast, a gossiping protocol. This algorithm has been designed to deal with computationally expensive problems via massive scalability; therefore, we analyse the response time of the model using large instances of well-known hard optimization problems that require from EAs a (sometimes exponentially) bigger computational effort as these problems scale. Our approach has been matched against a sequential Genetic Algorithm (sGA) applied to the same set of problems, and we found that it needs less computational effort than the sGA in yielding success. Furthermore, the response time scales logarithmically with respect to the problem size, which makes it suitable to tackle large instances.