Building and discovering useful features when constructing machine learning models is the central task for the machine learning practitioner. Good features are useful not only in increasing the predictive power of a model but also in illuminating the underlying drivers of a target variable. In this research we propose a novel feature learning technique in which Symbolic regression is endowed with a ``Range Terminal'' that allows it to explore functions of the aggregate of variables over time. We test the Range Terminal on a synthetic data set and a real world data in which we predict seasonal greenness using satellite derived temperature and snow data over a portion of the Arctic. On the synthetic data set we find Symbolic regression with the Range Terminal outperforms standard Symbolic regression and Lasso regression. On the Arctic data set we find it outperforms standard Symbolic regression, fails to beat the Lasso regression, but finds useful features describing the interaction between Land Surface Temperature, Snow, and seasonal vegetative growth in the Arctic.

Philosophiae Doctor - PhD / Promoter detection, especially in prokaryotes, has always been an uphill task and may remain so, because of the many varieties of sigma factors employed by various organisms in transcription. The situation is made more complex by the fact, that any seemingly unimportant sequence segment may be turned into a promoter sequence by an activator or repressor (if the actual promoter sequence is made unavailable). Nevertheless, a computational approach to promoter detection has to be performed due to number of reasons. The obvious that comes to mind is the long and tedious process involved in elucidating promoters in the ‘wet’ laboratories not to mention the financial aspect of such endeavors. Promoter detection/prediction of an organism with few characterized promoters (M.tuberculosis) as envisaged at the beginning of this work was never going to be easy. Even for the few known Mycobacterial promoters, most of the respective sigma factors associated with their transcription were not known. If the information (promoter-sigma) were available, the research would have been focused on categorizing the promoters according to sigma factors and training the methods on the respective categories. That is assuming that, there would be enough training data for the respective categories. Most promoter detection/prediction studies have been carried out on E.coli because of the availability of a number of experimentally characterized promoters (+- 310). Even then, no researcher to date has extended the research to the entire E.coli genome. / South Africa

With the advancement in multimedia systems and the increased interest in haptics to be used in interpersonal communication systems, where users can see, show, hear, tell, touch and be touched, mouse and keyboard are no longer dominant input devices. Touch, speech and vision will soon be the main methods of human computer interaction. Moreover, as interpersonal communication usage increases, the need for securing user authentication grows. In this research, we examine a user's identification and verification based on haptic information. We divide our research into three main steps. The first step is to examine a pre-defined task, namely a handwritten signature with haptic information. The user target in this task is to mimic the legitimate signature in order to be verified. As a second step, we consider the user's identification and verification based on user drawings. The user target is predefined, however there are no restrictions imposed on the order or on the level of details required for the drawing. Lastly, we examine the feasibility and possibility of distinguishing users based on their haptic interaction through an interpersonal communication system. In this third step, there are no restrictions on user movements, however a free movement to touch the remote party is expected. In order to achieve our goal, many classification and feature reduction techniques have been discovered and some new ones were proposed. Moreover, in this work we utilize evolutionary computing in user verification and identification. Analysis of haptic features and their significance on distinguishing users is hence examined. The results show a utilization of visual features by Genetic Programming (GP) towards identity verification, with a probability equal to 50% while the remaining haptic features were utilized with a probability of approximately 50%. Moreover, with a handwritten signature application, a verification success rate of 97.93% with False Acceptance Rate (FAR) of 1.28% and @11.54% False Rejection Rate (FRR) is achieved with the utilization of genetic programming enhanced with the random over sampled data set. In addition, with a totally free user movement in a haptic-enabled interpersonal communication system, an identification success rate of 83.3% is achieved when random forest classifier is utilized.

The motif discovery problem is becoming increasingly important for molecular biologists as new sequencing technologies are producing large amounts of data, at rates which are unprecedented. The solution space for DNA motifs is too large to search with naive methods, meaning there is a need for fast and accurate motif detection tools. We propose MotifGP, a multiobjective motif discovery tool evolving regular expressions that characterize overrepresented motifs in a given input dataset. This thesis describes and evaluates a multiobjective strongly typed genetic programming algorithm for the discovery of network expressions in DNA sequences. Using 13 realistic data sets, we compare the results of our tool, MotifGP, to that of DREME, a state-of-art program. MotifGP outperforms DREME when the motifs to be sought are long, and the specificity is distributed over the length of the motif. For shorter motifs, the performance of MotifGP compares favourably with the state-of-the-art method. Finally, we discuss the advantages of multi-objective optimization in the context of this specific motif discovery problem.

The main aim of this thesis is flow shop and job shop scheduling problem in logistics warehouses. Managing and scheduling works is currently often problem. There is no simple solution due to complexity of this problem. This problem must be resolved because of a lack efficiency of work with a higher load such as during the christmas holidays. This paper describes the methods used to solve this problem focusing mainly on the use of search algorithms, evolutionary algorithms, specifically grammar guided genetic programming. This paper describes the problem of job shop scheduling on a simple theoretical example. The implemented algorithm for solving this problem was subjected to tests inspired on data from real warehouse, as well as synthetically created tests with more jobs and a greater number of workers. Synthetic tests were generated randomly. All tests were therefore run several times and the results were averaged. In conclusion of this work are presented the results of the algorithm and the optimum parameter settings for different sizes of problems and requirements for the solution. Genetic algorithm has been extended to calculate fitness of individuals with regard to number of collisions, extended to use priority rules during run of evolution, and some parts of algorithm was parallelized.

The main goal of the work is to introduce new ideas how traditional approaches for designing an operation system and associated software can be improved to be a part of automatic software evolution. It is generally supposed that algorithms found by the genetic programming processes cannot be used for exact calculations but only for approximate solutions. Several examples of software evolution are introduced, to show that quite precise solutions can be achieved. To reach this goal, characteristics of tree-like structures with approaches based on cellular automata features are combined in a new promising technique of algorithm representation, joining benefits of both concepts. An application has been developed based on these new genetic programming concepts and it is supposed it can be a part of a future automatic software evolution process.

Optimization of digital circuits still attracts much attention not only of researchers but mainly chip producers. One of new the methods for the optimization of digital circuits is cartesian genetic programming. This Master's thesis describes a new crossover operator and its implementation for cartesian genetic programming. Experimental evaluation was performed in the task of three-bit multiplier and five-bit parity circuit design.

The diploma thesis deals with the evolutionary algorithms and their application in the area of digital circuit design. In the first part, general principles of evolutionary algorithms are introduced. This part includes also the introduction of genetic algorithms and genetic programming. The next chapter describes the cartesian genetic programming and its modifications like embedded, self-modifying or multi-chromosome cartessian genetic programming. Essential part of this work consists of the design and implementation of a modularization technique for evolution circuit design. The proposed approach is evaluated using a set of standard benchmark circuits.

Numerous optimization problems in production systems can be considered as decision-making processes that determine the best allocation of resources to tasks over time to optimize one or more objectives in concert with big data. Among the optimization problems, production scheduling and routing of robots for material handling are becoming more important due to their impacts on system performance. However, the development of efficient algorithms for scheduling or routing faces several challenges. While the scheduling and vehicle routing problems can be solved by mathematical models such as mixed-integer linear programming to find optimal solutions to smallsized problems, they are not applicable to larger problems due to the nature of NP-hard problems. Thus, further research on machine learning applications to those problems is a significant step towards increasing the possibilities and potentialities of field application. In order to create truly intelligent systems, new frameworks for scheduling and routing are proposed to utilize machine learning (ML) techniques. First, the dynamic single-machine scheduling problem for minimization of total weighted tardiness is addressed. In order to solve the problem more efficiently, a decisiontree-based approach called Generation of Rules Automatically with Feature construction and Treebased learning (GRAFT) is designed to extract dispatching rules from existing or good schedules. In addition to the single-machine scheduling problem, the flexible job-shop scheduling problem with release times for minimizing the total weighted tardiness is analyzed. As a ML-based solution approach, a random-forest-based approach called Random Forest for Obtaining Rules for Scheduling (RANFORS) is developed to solve the problem by generating dispatching rules automatically. Finally, an optimization problem for routing of autonomous robots for minimizing total tardiness of transportation requests is analyzed by decomposing it into three sub-problems. In order to solve the sub-problems, a comprehensive framework with consideration of conflicts between routes is proposed. Especially to the sub-problem for vehicle routing, a new local search algorithm called COntextual-Bandit-based Adaptive Local search with Tree-based regression (COBALT) that incorporates the contextual bandit into operator selection is developed. The findings from my research contribute to suggesting a guidance to practitioners for the applications of ML to scheduling and control problems, and ultimately to lead the implementation of smart factories.

When classifying high-dimensional data, a lot can be gained, in terms of both computational time and precision, by only considering the most important features. Many feature selection methods are based on the assumption that important features are highly correlated with their corresponding classes, but mainly uncorrelated with each other. Often, this assumption can help eliminate redundancies and produce good predictors using only a small subset of features. However, when the predictability depends on interactions between the features, such methods will fail to produce satisfactory results. Also, since the suitability of the selected features depends on the learning algorithm in which they will be used, correlation-based filter methods might not be optimal when using genetic programs as the final classifiers, as they fail to capture the possibly complex relationships that are expressible by the genetic programming rules. In this thesis a method that can find important features, both independently and dependently discriminative, is introduced. This method works by performing two different types of permutation tests that classifies each of the features as either irrelevant, independently predictive or dependently predictive. The proposed method directly evaluates the suitability of the features with respect to the learning algorithm in question. Also, in contrast to computationally expensive wrapper methods that require several subsets of features to be evaluated, a feature classification can be obtained after only one single pass, even though the time required does equal the training time of the classifier. The evaluation shows that the attributes chosen by the permutation tests always yield a classifier at least as good as the one obtained when all attributes are used during training - and often better. The proposed method also fares well when compared to other attribute selection methods such as RELIEFF and CFS. / Då man handskas med data av hög dimensionalitet kan man uppnå både bättre precision och förkortad exekveringstid genom att enbart fokusera på de viktigaste attributen. Många metoder för att hitta viktiga attribut är baserade på ett grundantagande om en stark korrelation mellan de viktiga attributen och dess tillhörande klass, men ofta även på ett oberoende mellan de individuella attributen. Detta kan å ena sidan leda till att överflödiga attribut lätt kan elimineras och därmed underlätta processen att hitta en bra klassifierare, men å andra sidan också ge missvisande resultat ifall förmågan att separera klasser i hög grad beror på interaktioner mellan olika attribut. Då lämpligheten av de valda attributen också beror på inlärningsalgoritmen i fråga är det troligtvis inte optimalt att använda sig av metoder som är baserade på korrelationer mellan individuella attribut och dess tillhörande klass, ifall målet är att skapa klassifierare i form av genetiska program, då sådana metoder troligtvis inte har förmågan att fånga de komplexa interaktioner som genetiska program faktiskt möjliggör. Det här arbetet introducerar en metod för att hitta viktiga attribut - både de som kan klassifiera data relativt oberoende och de som får sina krafter endast genom att utnyttja beroenden av andra attribut. Den föreslagna metoden baserar sig på två olika typer av permutationstester, där attribut permuteras mellan de olika dataexemplaren för att sedan klassifieras som antingen oberende, beroende eller irrelevanta. Lämpligheten av ett attribut utvärderas direkt med hänsyn till den valda inlärningsalgoritmen till skillnad från så kallade wrappers, som är tidskrävande då de kräver att flera delmängder av attribut utvärderas. Resultaten visar att de attribut som ansetts viktiga efter permutationstesten genererar klassifierare som är åtminstone lika bra som när alla attribut används, men ofta bättre. Metoden står sig också bra när den jämförs med andra metoder som till exempel RELIEFF och CFS.