Most classical unconstrained optimization methods require derivative information. Different methods have been proposed for problems where derivative information cannot be used. One class of these methods is heuristics including Evolutionary Algorithms (EAs). In this study, we propose EAs for unconstrained optimization under both deterministic and stochastic environments. We design a crossover operator that tries to lead the algorithm towards the global optimum even when the starting solutions are far from the optimal solution. We also adapt this algorithm to a stochastic environment where there exist only estimates for the function values. We design new parent selection schemes based on statistical grouping methods and a replacement scheme considering existing statistical information. We test the performance of our algorithms using functions from the literature and newly introduced functions and obtain promising results.

Imitation learning has been studied from a large range of disciplines, including adaptive robotics. In adaptive robotics the focus is often on how robots can learn tasks by imitating experts. In order to build robots able to imitate a number of problems must be solved, including: How does the robot know when and what to imitate? How does the robot link the recognition of observed actions to the execution of the same actions? This thesis presents an approach using unsupervised imitation where artificial evolution is used to find solutions to the problems. The approach is tested in a number of experiments where robots are being evolved to solve a number of navigation tasks of varying difficulty. Two sets of experiments are made for each task. In the first set the robots are trained without any demonstrator present. The second set is identical to the first one except for the presence of a demonstrator. The demonstrator is present in the beginning of the training and thereafter removed. The robots are not being programmed to imitate the demonstrator but are only instructed to solve the navigation tasks. By comparing the performance of the robots of the two sets the impact of the demonstrator is investigated. The results show that the robots evolved with a demonstrator need less training time than the robots evolved without any demonstrator except when the task is easy to solve in which case the demonstrator seems to have no effect on the performance of the robots. It is concluded that evolved robots are able to imitate demonstrators even if the robots are not explicitly programmed to follow the demonstrators.

The aim of inverting artificial neural networks (ANNs) is to find input patterns that are strongly classified as a predefined class. In this project an ANN is inverted by an evolutionary algorithm. The network is retrained by using the patterns extracted by the inversion as counter-examples, i.e. to classify the patterns as belonging to no class, which is the opposite of what the network previously did. The hypothesis is that the counter-examples extracted by the inversion will cause larger updates of the weights of the ANN and create a better mapping than what is caused by retraining using randomly generated counter-examples. This hypothesis is tested on recognition of pictures of handwritten digits. The tests indicate that this hypothesis is correct. However, the test- and training errors are higher when retraining using counter-examples, than for training only on examples of clean digits. It can be concluded that the counter-examples generated by the inversion have a great impact on the network. It is still unclear whether the quality of the network can be improved using this method.

Evolutionary algorithms are often used for hard optimization problems. Solving time of this problems is long, so we want effective parallelization for this algorithms. Unfortunately, classical methods of parallelization do not work very well in cases where the individual evaluations of problems take significantly different times. In this project, we will try to extend the evolu- tionary algorithm with interleaving generations, which offers a better use of computational resources than classical parallel evolutionary algorithms, by speculative evaluation. Speculative evaluation means the estimation of an in- dividual's fitness function and the prediction of the following steps, which we will use later in the case of a correct estimate. We compare the algorithm with speculative evaluation with the original version in a series of experi- ments and we look at the effect of accuracy in the speculative step on the performance of the algorithm. 1

In this article, a design methodology for complex composite aircraft structures is presented. The developed approach combines a multi-objective optimization method and a parameterized simulation model using a design concept database. Due to the combination of discrete and continuous design variables describing the structures, evolutionary algorithms are used within the presented optimization approach. The approach requires an evaluation of the design alternatives that is performed by parameterized simulation models. The variability of these models is achieved using a design concept database that contains different layouts for each implemented structural part. Due to the complexity of the generated aircraft structures, the finite element method is applied for the calculation of the structural behaviour. The applicability of the developed design approach will be demonstrated by optimizing two composite aircraft fuselage examples. The obtained results show that the developed methodology is useful and reliable for designing complex aircraft structures.

Despite the importance of power systems in today’s societies, they suffer from aging infrastructure and need to improve the efficiency, reliability, and security. Two issues that significantly limit the current grid’s efficient energy delivery and consumption are: loadollowing generation dispatch, and energy theft. A loadollowing generation dispatch is usually employed in power systems, which makes continuous small changes so as to account for differences between the actual energy demand and the predicted values. This approach has led to an average utilization of energy generation capacity below 55% [49]. Moreover, energy theft causes several billion dollar losses to U.S. utility companies [31] [16], while in developing countries it can amount to 50% of the total energy delivered [48]. Recently, the Smart Grid has been proposed as a new electric grid to modernize current power grids and enhance its efficiency, reliability, and sustainability. Particularly, in the Smart Grid, a digital communication network is deployed to enable two-way communications between users and system operators. It thus makes it possible to shape the users’ load demand curves by means of demand response strategies. Additionally, in the Smart Grid, traditional meters will be replaced with cyber-physical devices, called smart meters, capable of recording and transmitting users’ real-time power consumption. Due to their monitoring capabilities, smart meters offer a great opportunity to detect energy theft in smart grids, but also raise serious concerns about users’ privacy. In this dissertation, we design optimal load scheduling schemes to enhance system efficiency, and develop energy theft detection algorithms that can preserve users’ privacy.

No description available.

No description available.

With the rise in the application of evolution strategies for simulation optimization, a better understanding of how these algorithms are affected by the stochastic output produced by simulation models is needed. At very high levels of stochastic variance in the output, evolution strategies in their standard form experience difficulty locating the optimum. The degradation of the performance of evolution strategies in the presence of very high levels of variation can be attributed to the decrease in the proportion of correctly selected solutions as parents from which offspring solutions are generated. The proportion of solutions correctly selected as parents can be increased by conducting additional replications for each solution. However, experimental evaluation suggests that a very high proportion of correctly selected solutions as parents is not required. A proportion of correctly selected solutions of around 0.75 seems sufficient for evolution strategies to perform adequately. Integrating statistical techniques into the algorithm?s selection process does help evolution strategies cope with high levels of noise. There are four categories of techniques: statistical ranking and selection techniques, multiple comparison procedures, clustering techniques, and other techniques. Experimental comparison of indifference zone selection procedure by Dudewicz and Dalal (1975), sequential procedure by Kim and Nelson (2001), Tukey?s Procedure, clustering procedure by Calsinki and Corsten (1985), and Scheffe?s procedure (1985) under similar conditions suggests that the sequential ranking and selection procedure by Kim and Nelson (2001) helps evolution strategies cope with noise using the smallest number of replications. However, all of the techniques required a rather large number of replications, which suggests that better methods are needed. Experimental results also indicate that a statistical procedure is especially required during the later generations when solutions are spaced closely together in the search space (response surface).

Tabletop games have within recent years evolvedto become more and more complex, such as through the useof dynamic rules, permanently changing how the game worksafter a playthrough, and players playing different roles in thegame. This leads to unique challenges for Artificial Intelligence.A Tabletop Games Framework (TAG) is a framework intended topromote research within general AI for modern tabletop games.Rolling Horizon Evolutionary Algorithms (RHEA) are a typeof algorithms that have been applied to games with successin the past. By implementing a RHEA agent we can studyhow it compares to other types of agents such as Monte CarloTree Search and Random Mutation Hill Climbing agents. Ofparticular interest is the game Pandemic (2008), as the existingagents are unable to win at it.