Sweden experiences heavy snowfall during the winter season and cost effective road maintenance is significantly affected by the routing of snowplows. The routing problem becomes more complex as the SwedishNational Road Administration (Vägverket) sets operational requirements such as satisfying a time window for each road segment. This thesis focuses on route optimization for snowplows after snowfall; to develop and implement an algorithm for finding combinations of generated routes which minimize the total cost. The results are compared to those stated in the licentiate thesis by Doctoral student Nima Golbaharan (2001). The algorithm calculates a lower bound to the problem using a Lagrangian master problem. A common subgradient approach is used to find near-optimal dual variables to be sent to a column-generation program which returns routes for the snowplows. A greedy heuristic chooses a feasible solution, which gives an upper bound to the problem. This entire process is repeated as needed. This method for routing snowplows produces favorable results with a relatively small number of routes and are comparable to Golbaharan's results. An interesting observation involves the allocation of vehicles in which certain depots were regularly over- or under-utilized. This suggests that the quantity and/or distribution of available vehicles may not be optimal.

Nynas produces bitumen at two refineries in Sweden. The bitumen is shipped to seven depots along the swedish coast line, and from the depots special trucks handle the transportation to customers. Recently Nynas has transformed its supply chain and closed down a few depots. At the moment the company is considering a further reduction of the number of depots. In connection to these discussions an analyse of the companys distributionsystem and of possible changes is required. In this thesis an optimization model is developed that simulates Nynas distribution of bitumen from the depots to the customers. The model is used to investigate the required vehicle fleet size for a number of different scenarios, that is with different depots closed down. The question to be answered is, thus, what depots could be closed without any dramatic increase in the required vehicle fleet size? Scenarios where customers are allocated an increased storage capacity are also studied. The distribution model that is developed is an inventory route planning problem. It is solved by column generation. Each column represents a route and is generated by a subproblem with restrictions on permitted working hours for the truck drivers. Integer solutions are generated heuristically. Simulations that have been performed with the model reveals interesting differences concerning how the distribution is handled in different parts of Sweden. In western Sweden the transportation planning works well, but the distribution in the central parts of the country could be planned in a better way. Results from simulations also show that the depots in Norrköping and Västerås could be closed down without increasing the vehicle fleet. Probably, the existing vehicle fleet size will be sufficient even with the Kalmar-depot closed down. Nevertheless, Nynas transportation suppliers will have to purchase new vehicles if the Sandarne-depot is to be closed. Another interesting conclusion that can be drawn from this thesis is that there is a potential for reducing the vehicle fleet size if the storage capacity is increased at a few chosen customers. A considerably small increase in the storage capacity at a few big customers that are located far from the depots will have a great effect.

In this thesis, the basic principles and concepts of single and multi-objective Genetic Algorithms (GA) are reviewed. Two algorithms, one for single objective and the other for multi-objective problems, which are believed to be more efficient are described in details. The algorithms are coded with MATLAB and applied on several test functions. The results are compared with the existing solutions in literatures and shows promising results. Obtained pareto-fronts are exactly similar to the true pareto-fronts with a good spread of solution throughout the optimal region. Constraint handling techniques are studied and applied in the two algorithms. Constrained benchmarks are optimized and the outcomes show the ability of algorithm in maintaining solutions in the entire pareto-optimal region. In the end, a hybrid method based on the combination of the two algorithms is introduced and the performance is discussed. It is concluded that no significant strength is observed within the approach and more research is required on this topic. For further investigation on the performance of the proposed techniques, implementation on real-world engineering applications are recommended.

Applications of the existing steady-state plant-wide optimization and the single-scale fast-rate dynamic optimization strategies to an integrated plant with material recycle have been impeded by several factors. While the steady-state optimization formulation is very simple, the very long transient dynamics of an integrated plant have limited the optimizers execution rate to be extremely low, yielding a suboptimal performance. In contrast, performing dynamic plant-wide optimization at the same rate as local controllers requires exorbitant on-line computational load and may increase the sensitivity to high-frequency dynamics that are irrelevant to the plant-level interactions, which are slow-scale in nature. This thesis proposes a novel multi-scale dynamic optimization and control strategy suitable for an integrated plant. The dynamic plant-wide optimizer in this framework executes at a slow rate to track the slow-scale plant-wide interactions and economics, while leaving the local controllers to handle fast changes related to the local units. Moreover, this slow execution rate demands less computational and modeling requirement than the fast-rate optimizer. An important issue of this method is obtaining a suitable dynamic model when first-principles are unavailable. The difficulties in the system identification process are designing proper input signal to excite this ill-conditioned system and handling the lack of slow-scale dynamic data when the plant experiment cannot be conducted for a long time compared to the settling time. This work presents a grey-box modeling method to incorporate steady-state information to improve the model prediction accuracy. A case study of an integrated plant example is presented to address limitations of the nonlinear model predictive control (NMPC) in terms of the on-line computation and its inability to handle stochastic uncertainties. Then, the approximate dynamic programming (ADP) framework is investigated. This method computes an optimal operating policy under uncertainties off-line. Then, the on-line multi-stage optimization can be transformed into a single-stage problem, thus reducing the real-time computational effort drastically. However, the existing ADP framework is not suitable for an integrated plant with high dimensional state and action space. In this study, we combine several techniques with ADP to apply nonlinear optimal control to the integrated plant example and show its efficacy over NMPC.

The objective of this study is to investigate the use and efficiency of Differential Evolution (DE) method on structural optimization. The solution algorithm developed with DE is computerized into software called SOP2011 using VB.NET. SOP2011 is automated to achieve size optimum design of steel structures consisting of 1-D elements such as trusses and frames subjected to design provisions according to ASD-AISC (2010) and LRFD-AISC (2010). SOP2011 works simultaneously with the structural analysis and design software SAP2000 in order to find the global or near optimum designs for real size truss and frame structures in which the optimization problem is classified as constrained, discrete size optimization. Software interacts with SAP2000 through the Open Application Programming Interface (OAPI), which provides an access to information of SAP2000 inputs and outputs. It is programmed for finding reasonable and optimized results for truss and frame steel structures by choosing appropriate ready sections for structural members considering the minimum weight via DE technique. Based on the comparison of the obtained results with the literature, DE algorithm with penalty function implementation is proved to be an efficient optimization technique amongst several major methods used for discrete constrained size optimization of real size steel structures. Also, it has been shown that by using optimized designs obtained by DE, weight of the structures can be reduced up to 67.9% for steel truss structures and 41.7% for steel frame structures compared to SAP2000 auto design procedure, hence resulting a significant saving of materials, cost, work hours and energy required for the project.

A Bayesian adaptive sampling method is developed for highly coupled multidisciplinary design problems. The method addresses a major challenge in aerospace design: exploration of a design space with computationally expensive analysis tools such as computational fluid dynamics (CFD) or finite element analysis. With a limited analysis budget, it is often impossible to optimize directly or to explore a design space with off-line design of experiments (DoE) and surrogate models. This difficulty is magnified in multidisciplinary problems with feedbacks between disciplines because each design point may require iterative analyses to converge on a compatible solution between different disciplines. Bayesian Collaborative Sampling (BCS) is a bi-level architecture for adaptive sampling that simulataneously - concentrates disciplinary analyses in regions of a design space that are favorable to a system-level objective - guides analyses to regions where interdisciplinary coupling variables are probably compatible BCS uses Bayesian models and sequential sampling techniques along with elements of the collaborative optimization (CO) architecture for multidisciplinary optimization. The method is tested with the aero-structural design of a glider wing and the aero-propulsion design of a turbojet engine nacelle.

Accurately representing the price-demand relationship is critical for the success of a price optimization system. This research first uses booking data from 28 U.S. hotels to investigate the validity of two key assumptions in hotel revenue management. The assumptions are: 1) customers who book later are willing to pay higher rates than customers who book earlier; and, 2) demand is stronger during the week than on the weekend. Empirical results based on an analysis of booking curves, average paid rates, and occupancy rates for group, restricted retail, unrestricted retail, and negotiated demand segments challenge the validity of these assumptions. The combination of lower utilization rates and greater product differentiation suggests that hotels should apply different approaches than simply matching competitor rates to avoid losing market share. On days when inventory is near capacity, traditional yield management tactics deliver tremendous value, but these should be augmented by incorporating price response of demand and competition effects. On days when demand is soft and occupancy is projected to be low, price and competition based strategies should dominate. The hotel price optimization problem with linear demand model is a quadratic programming problem with prices of products that utilize multiple staynight rooms as the decision variable. The optimal solution of the hotel price optimization problems has unique properties that enables us to develop an alternative optimization algorithm that does not require solving quadratic optimization problem. Using the well known least norm problem as a subroutine, the optimization problem can be solved as finding a minimum distance between a polyhedron defined by non-negative demand and capacity constraints. This algorithm is efficient when only a few of the staynights are highly constrained. In practice, the choice of a demand model is largely driven by the ease of estimation and model fit statistics such as R2 and mean absolute percentage error (MAPE). These metrics provide measures of statistical validity of the model, however, they do not measure how well the price optimization will perform which is the ultimate interest of the practitioners. In order to measure the impact of demand models on price optimization performance, we first investigate the goodness of fit of linear demand models with different driver variables using actual data from 23 U.S. hotels representing multiple brands and location types. We find that hotels within the same location types (such as urban, suburban, airport) share similar driver variables. Airport and suburban hotels have simpler model specifications with less drivers compared to the urban hotels. The airport hotel demand models are different from other location hotels in that the airport hotel demand level does not differ by day of week. We then measure the impact of demand model misrepresentation on the performance of price optimization through simulation experiments, which are performed for different levels of demand and forecast accuracy to represent various market environments that hotels operate in. We find that using models with missing driver variables can reduce the potential revenue by 13%∼53% and using the wrong functional form 5%∼43% under our simulation environment. The findings from our research imply that correctly representing the demand model in price optimization is crucial to its success. In order for hotels to realize the maximum potential revenue through pricing, efforts should be focused on identifying the major driver variables influencing demand including the ones that we found to be significant.

Simulation-based optimization methodologies are widely applied in real world optimization problems. In developing these methodologies, beside simulation models, algorithms play a critical role. One example is an evolutionary multi objective optimization algorithm known as Reference point-based Non-dominated Sorting Genetic Algorithm-II (R-NSGA-II), which has shown to have some promising results in this regard. Its successor, R-NSGA-II-adaptive diversity control (hereafter Adaptive Epsilon-R-NSGA-II (AE-R-NSGA-II) algorithm) is one of the latest proposed extensions of the R-NSGA-II algorithm and in the early stages of its development. So far, little research exists on its applicability and usefulness, especially in real world optimization problems. This thesis evaluates behavior and performance of AE-R-NSGA-II, and to the best of our knowledge is one of its kind. To this aim, we have investigated the algorithm in two experiments, using two benchmark functions, 10 performance measures, and a behavioral characteristics analysis method. The experiments are designed to (i) assess behavior and performance of AE-R-NSGA-II, (ii) and facilitate efficient use of the algorithm in real world optimization problems. This is achieved through the algorithm parameter configuration (parametric study) according to the problem characteristics. The behavior and performance of the algorithm in terms of diversity of the solutions obtained, and their convergence to the optimal Pareto front is studied in the first experiment through manipulating a parameter of the algorithm referred to as Adaptive epsilon coefficient value (C), and in the second experiment through manipulating the Reference point (R) according to the distance between the reference point and the global Pareto front. Therefore, as one contribution of this study two new diversity performance measures (called Modified spread, and Population diversity), and the behavioral characteristics analysis method called R-NSGA-II adaptive epsilon value have been introduced and applied. They can be modified and applied for the evaluation of any reference point based algorithm such as the AE-R-NSGA-II. Additionally, this project contributed to improving the Benchmark software, for instance by identifying new features that can facilitate future research in this area. Some of the findings of the study are as follows: (i) systematic changes of C and R parameters influence the diversity and convergence of the obtained solutions (to the optimal Pareto front and to the reference point), (ii) there is a tradeoff between the diversity and convergence speed, according to the systematic changes in the settings, (iii) the proposed diversity measures and the method are applicable and useful in combination with other performance measures. Moreover, we realized that because of the unexpected abnormal behaviors of the algorithm, in some cases the results are conflicting, therefore, impossible to interpret. This shows that still further research is required to verify the applicability and usefulness of AE-R-NSGA-II in practice. The knowledge gained in this study helps improving the algorithm.

O projeto de um componente mecânico é uma atividade muito complexa, onde muitas vezes se tem restrições de projeto como peso do componente e rigidez máxima, e também restrições de manufatura, associada aos processos de fabricação disponíveis para serem utilizados. É fato conhecido que a Otimização Topológica (OT), apesar de ser um método extremamente eficiente para a obtenção de soluções ótimas, gera soluções com geometrias complexas que são ou muito caras de se fabricar ou infactíveis. A técnica de projeção foi escolhida como adequada para implementar as restrições propostas neste trabalho. Esta técnica resolve o problema posto num domínio de variáveis de projeto e projeta essa solução num domínio de pseudo-densidades, que são a resposta do problema. A relação entre os dois domínios e determinada pela função de projeção e pelo mapeamento das variáveis definidos de forma diferente para cada restrição. Neste trabalho foram implementadas restrições de manufatura para OT de modo a restringir a gama possível de soluções no problema de otimização. Como exemplo foi considerado o problema de maximização de rigidez, com restrição de volume. Todas as implementações foram realizadas em linguagem de programação C, e o algoritmo de otimização utilizado é o critério de optimalidade. Foram implementadas as seguintes restrições de manufatura com a técnica de projeção: membro mínimo, buraco mínimo, simetria, extrusão, é revolução, repetição de padrões, fundição, forjamento, e laminação. Estas restrições mostram a grande capacidade da técnica de projeção para controlar a solução do problema de otimização sem implicar num grande aumento do custo computacional. Os resultados encontrados mostram a potencialidade de utilizar restrições de manufatura na OT, porém estão longe de esgotarem o assunto, nesse tema recente que vem sendo explorado no Método de Otimização Topológica (MOT). / The design of a mechanical component is a very complex task, which includes constraints such as maximum weight and maximum stiffness, and also manufacturing constraints, associated with the manufacturing processes required at the shop floor. It is known that Topology Optimization (TO), despite of being a very effective and powerful method to obtain optimal solutions, generates solutions with complex geometries that are too much expensive to be manufactured or just can not be made. The projection scheme has been chosen as the most appropriate technique for implementing the proposed constraints. This scheme solves the proposed problem in a domain of design variables and then projects these results into a pseudo-density domain to find the solution. The relation between both domains is defined by the projection function and variable mapping defined in a different way for each constraint. In this work, manufacturing constraints for TO are implemented in a way that the possible solutions of the optimization problem are restricted. As an example, the traditional stiffness maximization problem is considered. All implementations have been done using C programming language, and the optimization algorithm applied is the optimality criteria. The following manufacturing constraints have been implemented using the projection scheme: minimal member size, minimal hole size, symmetry, extrusion, revolution, pattern repetition, casting, forging and lamination. These constraints show the large capacity of the projection scheme to control the solution for the optimization without adding a large computational cost. The results that have been found show the great power of using manufacturing constraints in the TO, however, they are far from exhausting this topic that has been recently explored in the Topology Optimization Method (TOM).

Orientadores: Fernando Jose Von Zuben, Leandro Nunes de Castro / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-14T19:14:33Z (GMT). No. of bitstreams: 1 Franca_FabricioOlivettide_M.pdf: 2824607 bytes, checksum: 3de6277fbb2c8c3460d62b4d81d14f73 (MD5) Previous issue date: 2005 / Resumo: Esta dissertação propõe algoritmos bio-inspirados para a solução de problemas de otimização dinâmica, ou seja, problemas em que a superfície de otimização no espaço de busca sofre variações diversas ao longo do tempo. Com a variação, no tempo, de número, posição e qualidade dos ótimos locais, as técnicas de programação matemática tendem a apresentar uma acentuada degradação de desempenho, pois geralmente foram concebidas para tratar do caso estático. Algoritmos populacionais, controle dinâmico do número de indivíduos na população, estratégias de busca local e uso eficaz de memória são requisitos desejados para o sucesso da otimização dinâmica, sendo contemplados nas propostas de solução implementadas nesta dissertação. Os algoritmos a serem apresentados e comparados com alternativas competitivas presentes na literatura são baseados em funcionalidades e estruturas de processamento de sistemas imunológicos e de colônias de formigas. Pelo fato de considerarem todos os requisitos para uma busca eficaz em ambientes dinâmicos, o desempenho dos algoritmos imuno-inspirados se mostrou superior em todos os critérios considerados para comparação dos resultados dos experimentos. / Abstract: This dissertation proposes bio-inspired algorithms to solve dynamic optimization problems, i.e., problems for which the optimization surface on the search space suffers several changes over time. With such variation of number, position and quality of local optima, mathematical programming techniques may present degradation of performance, because they were usually conceived to deal with static problems. Population-based algorithms, dynamic control of the population size, local search strategies and an efficient memory usage are desirable requirements to a proper treatment of dynamic optimization problems, thus being incorporated into the solution strategies implemented here. The algorithms to be presented, and compared with competitive alternatives available in the literature, are based on functionalities and processing structures of immune systems and ant colonies. Due to the capability of incorporating all the requirements for an efficient search on dynamic environments, the immune-inspired approaches overcome the others in all the performance criteria adopted to evaluate the experimental results. / Mestrado / Engenharia de Computação / Mestre em Engenharia Elétrica