Recent advances in the field of computational intelligence have led to a numberof promising optimization algorithms. These algorithms have the potential to findoptimal or near-optimal solutions to complex problems within a reasonable timeframe. Structural optimization is a research field where such algorithms are appliedto optimally design structures. Although a significant amount of research has been published in the field ofstructural optimization since the 1960s, little of the research effort has been utilizedin structural design practice. One reason for this is that only a small portion ofthe research targets real-world applications. Therefore there is a need to conductresearch on cost optimization of realistic structures, particularly large structureswhere significant cost savings may be possible. To address this need, a software application for cost optimization of beam bridgeswas developed. The software application was limited to road bridges in concretethat are straight and has a constant width of the bridge deck. Several simplificationswere also made to limit the scope of the thesis. For instance, a rough design ofthe substructure was implemented, and the design of some structural parts wereneglected. This thesis introduces the subject of cost optimization, treats fundamentaloptimization theory, explains how the software application works, and presents acase study that was carried out to evaluate the application. The result of the case study suggests a potential for significant cost savings. Yet,the speeding up of the design process is perhaps the major benefit that should inclinedesigners to favor optimization. These findings mean that current optimizationalgorithms are robust enough to decrease the cost of beam bridges compared to aconventional design. However, the software application needs several improvementsbefore it can be used in a real design situation, which is a topic for future research. / Nya framsteg inom forskningen har lett till ett antal lovande optimeringsalgoritmer.Dessa algoritmer har potentialen att hitta optimala eller nästan optimala lösningartill komplexa problem inom rimlig tid. Strukturoptimering är ett forskningsområdedär dessa algoritmer tillämpas för att dimensionera konstruktioner på ett optimaltsätt. Även om en betydande mängd forskning har publicerats inom området strukturoptimeringsedan 1960-talet, så har endast lite av forskningsinsatserna kommit tillanvändning i praktiken. Ett skäl till detta är att endast en liten del av forskningenär inriktad mot verklighetsförankrade tillämpningar. Därför finns det ett behov avatt bedriva forskning på kostnadsoptimering av realistiska konstruktioner, särskiltstora konstruktioner där betydande kostnadsbesparingar kan vara möjligt. För att möta detta behov har ett datorprogram för kostnadsoptimering avbalkbroar utvecklats. Programmet begränsades till vägbroar i betong som är rakaoch har en konstant bredd. Flera förenklingar gjordes också för att begränsaomfattningen av arbetet. Till exempel implementerades en grov dimensionering avunderbyggnaden, och dimensioneringen av vissa komponenter försummades helt ochhållet. Detta examensarbete presenterar ämnet kostnadsoptimering, behandlar grundläggandeoptimeringsteori, förklarar hur programmet fungerar, och presenterar enfallstudie som genomfördes för att utvärdera programmet. Resultatet av fallstudien visar en potential för betydande kostnadsbesparingar.Trots det så är tidsbesparingarna i dimensioneringsprocessen kanske den störstafördelen som borde locka konstruktörer att använda optimering. Dessa upptäckterinnebär att aktuella optimeringsalgoritmer är tillräckligt robusta för att minskakostnaden för balkbroar jämfört med en konventionell dimensionering. Dock måsteprogrammet förbättras på flera punkter innan det kan användas i en verklig dimensioneringssituation,vilket är ett ämne för framtida forskning.

Manual design of levels can be an expensive and time consuming process. Procedural content generation (PCG) entails methods to algorithmically generate game content such as levels. One such way is by using cellular automata (CA), and in particular evolved cellular automata. Existing research primarily considers specifically determined starting states, as opposed to randomly initialized ones. In this paper we investigate the current state of the art regarding using CA’s that have been evolved with a genetic algorithm (GA) for level generation purposes. Additionally, we create a level generator that uses a GA in order to evolve CA rules for the creation of maze-like 2d levels which can be used in video games. Specifically, we investigate if it is possible to evolve CA rules that, when applied to a set of random starting states, could transform these into game levels with long solution paths and a large number of dead ends. We generate 60 levels over 6 experiments, rendering 58 playable levels. Our analysis of the levels show some flaws in certain levels, such as large numbers of unreachable cells. Additionally, the results indicate that the designed GA can be further improved upon. Finally, we conclude that it is possible to evolve CA rules that can transform a set of random starting states into game levels.

No description available.

Metaheuristics is a term for optimization procedures/algorithms that can be applied to a wide range of problems. These problems for which metaheuristics are used usually fall in the NP-hard category, meaning that they cannot be solved in polynomial time. This means that as the input dataset gets larger the time to solve increases exponentially. One such problem is the traveling salesman problem (TSP) which is and has been widely used as a benchmark problem to test optimization algorithms. This study focused on two such algorithms called ant colony optimization (ACO) and genetic algorithm (GA) respectively. Development of such optimization algorithms can have huge implications in several areas of business and industry. They can for example be used by delivery companies to optimize routing of delivery vehicles as well as in material science/industry where they can be used to calculate the most optimal mix of ingredients to produce materials with the desired characteristics. The approach taken in this study was to compare the performance of the two algorithms in three different programming languages (python, javascript and C#).  Previous studies comparing the two algorithms have reported conflicting results where some studies found that ACO yielded better results but was slower than GA, while others found that GA yielded better results than ACO. Results of this study suggested that both ACO and GA could find the benchmark solution, but  ACO did so much more consistently. Furthermore javascript was found to be the most efficient language with which to run the algorithms in the setup used in this study.

The search space of digital 3D models designs is vast and can be hard to navigate. In this study, a system that evolves digital 3D models using an interactive genetic algorithm (IGA) was constructed in order to aid this process. The goal of the study was to investigate how such a system can be constructed in order to aid the design space exploration of digital 3D models.  The system is integrated with the 3D creation suite Blender and uses its Python API to programmatically edit models and generate images of the results, which are displayed on a web page where users can rate the results to evolve the model. The proposed system exposes all settings for the genetic algorithm, which includes population size, mutation rate, crossover algorithm, selection algorithm and more. Furthermore, the settings can be modified throughout the evolutionary process as well as the ability to rewind the algorithm and go back to previous generations in order to give more control in the progression of the algorithm. The script based nature of the proposed system is powerful but not practical for people without programming experience. For widespread adoption of IGAs as an exploratory design aid tool, it would help if the IGA is directly integrated into the design software being used in order to make it easier to use and reduce user fatigue.

For decades, dengue virus has been a cause of major public health concern in Costa Rica, due to its landscape and climatic conditions that favor the circumstances in which the vector, Aedes aegypti, thrives. The emergence and introduction throughout tropical and subtropical countries of the chikungunya virus, as of 2014, challenged Costa Rican health authorities to provide a correct diagnosis since it is also transmitted by the same vector and infected hosts may share similar symptoms. We study the 2015–2016 dengue and chikungunya outbreaks in Costa Rica while establishing how point estimates of epidemic parameters for both diseases compare to one another. Longitudinal weekly incidence reports of these outbreaks signal likely misdiagnosis of infected individuals: underreporting of chikungunya cases, while overreporting cases of dengue. Our comparative analysis is formulated with a single-outbreak deterministic model that features an undiagnosed class. Additionally, we also used a genetic algorithm in the context of weighted least squares to calculate point estimates of key model parameters and initial conditions, while formally quantifying misdiagnosis.

Deep packet inspection (DPI) is useful as a tool for analyzing internet traffic. Regular expressions (regexps) can be used to detect the network traffic patterns that the DPI is able to identify. These regexps can be represented as state-machines, and sometimes combining smaller state-machines into larger state-machines can result in more efficient processing. This thesis looks at how to decide which state-machines used in DPI-classes should be combined with which other state-machines in an efficient manner using genetic algorithms. The goal being to create as few resulting state-machines from the combination while still maintaining a upper limit on the size of the resulting state-machines. The problem is modelled as an assignment problem for which an emulated surrogate problem is used in order to make experimental evaluations. Several genetic algorithms are suggested in an attempt to explore a wide range of parameters. It is also evaluated if different genetic algorithms perform differently depending on if the state-machines represent DPI-classes used to parse UDP or TCP traffic. A 2-dimensional representation is used that allows for a better capture of the underlying assignment problem. Different approaches to fitness are explored and found to have different efficacy in different situations. Several genetic algorithm operators are suggested for different situations and a difference is found between what works for UDP and for TCP. / Deep packet inspection (DPI) ̈ar användbart som ett verktyg f ̈or att analysera internettrafik. Regular expressions (regexps) kan användas för att detektera trafik mönster somDPI:n kan identifiera. De här regexps kan representeras som state-machines, och ibland så kan kombinationen av mindre state-machines till större state-machines resultera i mer effektiv bearbetning. Den här tesen undersöker hur man kan bestämma vilka state-machines som används iDPI-klassen bör bli kombinerade på ett effektivt sätt med genetiska algoritmer. Målet är att skapa så fǻ resulterande state-machines från kombineringen på ett sådant sätt att storleken på alla resulterande state-machines håller sig under en övre gräns. Problemet är modellerat som ett assignment problem för vilket ett emulerat surrogatproblem används för att tillåta experiment att utföras. Ett flertal genetiska algoritmer är föreslagna i ett försök att undersöka en bred räckvidd av parametrar. Det är också undersökt om olika genetiska algoritmer har olika prestanda beroende på om state-machines representerar DPI-klasser använda för UDP eller TCP trafik. En 2-dimensionell representation som fångar det underliggande problemet på ett bras sätt är använd. Olika tillvägagångssätt för att representera fitness är undersökta och är upptäckta att ha olika effektivitet i olika situationer. Ett flertal genetiska algoritm operatorer är föreslagna för olika situationer och en skillnad är hittad mellan vad som fungerar för UDP och vad som fungerar för TCP.

Many real world problems in science and engineering can be treated as optimization problems with multiple objectives or criteria. The demand for fast and robust stochastic algorithms to cater to the optimization needs is very high. When the cost function for the problem is nonlinear and non-differentiable, direct search approaches are the methods of choice. Many such approaches use the greedy criterion, which is based on accepting the new parameter vector only if it reduces the value of the cost function. This could result in fast convergence, but also in misconvergence where it could lead the vectors to get trapped in local minima. Inherently, parallel search techniques have more exploratory power. These techniques discourage premature convergence and consequently, there are some candidate solution vectors which do not converge to the global minimum solution at any point of time. Rather, they constantly explore the whole search space for other possible solutions. In this thesis, we concentrate on benchmarking three popular algorithms: Real-valued Genetic Algorithm (RGA), Particle Swarm Optimization (PSO), and Differential Evolution (DE). The DE algorithm is found to out-perform the other algorithms in fast convergence and in attaining low-cost function values. The DE algorithm is selected and used to build a model for forecasting auroral oval boundaries during a solar storm event. This is compared against an established model by Feldstein and Starkov. As an extended study, the ability of the DE is further put into test in another example of a nonlinear system study, by using it to study and design phase-locked loop circuits. In particular, the algorithm is used to obtain circuit parameters when frequency steps are applied at the input at particular instances.

Advanced Metering Infrastructure (AMI) is one of the most important components of smart grid (SG) which aggregates data from smart meters (SMs) and sends the collected data to the utility center (UC) to be analyzed and stored. In traditional centralized AMI architecture, there is one meter data management system to process all gathered information in the UC, therefore, by increasing the number of SMs and their data rates, this architecture is not scalable and able to satisfy SG requirements, e.g., delay and reliability. Since scalability is one of most important characteristics of AMI architecture in SG, we have investigated the scalability of different AMI architectures and proposed a scalable hybrid AMI architecture. We have introduced three performance metrics. Based on these metrics, we formulated each AMI architecture and used a genetic-based algorithm to minimize these metrics for the proposed architecture. We simulated different AMI architectures for five demographic regions and the results proved that our proposed AMI hybrid architecture has a better performance compared with centralized and decentralized AMI architectures and it has a good load and geographic scalability.

Indiana University-Purdue University Indianapolis (IUPUI) / Design optimization of composite structures is of importance in the automotive, aerospace, and energy industry. The majority of optimization methods applied to laminated composites consider linear or simplified nonlinear models. Also, various techniques lack the ability to consider the composite failure criteria. Using artificial neural networks approximates the objective function to make it possible to use other techniques to solve the optimization problem. The present work describes an optimization process used to find the optimum design to meet crashworthiness requirements which includes minimizing peak crushing force and specific energy absorption for a square tube. The design variables include the number of plies, ply angle and ply thickness of the square tube. To obtain an effective approximation an artificial neural network (ANN) is used. Training data for the artificial neural network is obtained by crash analysis of a square tube for various samples using LS DYNA. The sampling plan is created using Latin Hypercube Sampling. The square tube is considered to be impacted by the rigid wall with fixed velocity and rigid body acceleration, force versus displacement curves are plotted to obtain values for crushing force, deceleration, crush length and specific energy absorbed. The optimized values for the square tube to fulfill the crashworthiness requirements are obtained using an artificial neural network combined with Multi-Objective Genetic Algorithms (MOGA). MOGA finds optimum values in the feasible design space. Optimal solutions obtained are presented by the Pareto frontier curve. The optimization is performed with accuracy considering 5% error.