Spatial layout design must consider not only ease of movement for pedestrians under normal conditions, but also their safety in panic situations, such as an emergency evacuation in a theatre, stadium or hospital. Using pedestrian simulation statistics, the movement of crowds can be used to study the consequences of different spatial layouts. Previous works either create an optimal spatial arrangement or an optimal pedestrian circulation. They do not automatically optimise both problems simultaneously. Thus, the idea behind the research in this thesis is to achieve a vital architectural design goal by automatically producing an optimal spatial layout that will enable smooth pedestrian flow. The automated process developed here allows the rapid identification of layouts for large, complex, spatial layout problems. This is achieved by using Cellular Automata (CA) to model pedestrian simulation so that pedestrian flow can be explored at a microscopic level and designing a fitness function for heuristic search that maximises these pedestrian flow statistics in the CA simulation. An analysis of pedestrian flow statistics generated from feasible novel design solutions generated using the heuristic search techniques (hill climbing, simulated annealing and genetic algorithm style operators) is conducted. The statistics that are obtained from the pedestrian simulation is used to measure and analyse pedestrian flow behaviour. The analysis from the statistical results also provides the indication of the quality of the spatial layout design generated. The technique has shown promising results in finding acceptable solutions to this problem when incorporated with the pedestrian simulator when demonstrated on simulated and real-world layouts with real pedestrian data.

The research presented in this thesis investigates the performance of some well-known heuristic algorithms on graph decomposition problems. First, a genetic algorithm is introduced and some modifications are trialled on finding Steiner triple systems (STS) of small orders. The results show that traditional genetic algorithms are not well suited to finding graph decompositions. Then a hill climbing optimisation technique is presented and investigated in the context of cycle decompositions. Such searches have previously proved to be effective at finding STSs. However, the general hill climbing approach is not immediately applicable to decompositions into cycles of length larger than 3. A modification of the hill climbing algorithm for cycles, called slippery hill climbing, is introduced and tested on decompositions of graphs into cycles of small lengths larger than 3. Slippery hill climbing successfully decomposed complete and dense non-complete graphs of considerable sizes into cycles of small lengths. In addition, we applied the slippery hill climbing approach to completing partial latin squares. It is reasonably expected that the algorithms developed in this study will also be applicable to other related problems in combinatorics and graph theory.

Konstanz, Univ., Diplomarb., 2005.

Airships provide an eco-friendly and cost-effective means to suit sustained airborne operations. Smaller autonomous airships are highly susceptible to adverse atmospheric conditions owing to their under-actuated, underpowered and bulky size relative to other types of unmanned aerial vehicles (UAVs). To mitigate these limitations, careful considerations of the size and shape must be made at the design stage. This research presents a methodology for obtaining an optimized shape of a semi-rigid airship. Rapid descent of the LTA ship is achieved by means of a moving gondola attached to a rigid keel mounted under the helium envelope from the bow to the mid-section of the hull. The study entails the application of a robust hybrid genetic algorithm (HGA) for the multi-disciplinary design and optimization of an airship capable of rapid descent, with lower drag and optimum surface area. A comprehensive sensitivity analysis was also performed on the basis of algorithmic parameters and atmospheric conditions. With the help of HGA, a semi-rigid airship capable of carrying a payload of 0.25 kg to 1.0 kg and capable of pitching at right angles is conceptually designed. The algorithm is also tested on commercially available vehicles to validate the results. In multi-objective optimization problems (MOOPs), the significance of different objectives is dependent on the user.

This paper researches the impact of different settings on an Infinite Impulse Response-filter (IIR-filter) used on a NIBE heat pump in combination with photovoltaic panels (PV-panel). The IIR-filter is applied to the level of the PV-panel’s power and its output is used by the heat pump’s control to harvest as much solar power as possible for supplying the heat pump with electricity. In some of the experiments weather data is used in the form of a forecast regarding the incoming cloudiness in the area, called “cloud coverage”. My objective is to find out which setting performs the best, and whether an evolutionary algorithm can find an optimal setting. The evolutionary algorithms I try are Genetic Algorithm, Simulated Annealing and the Hill Climbing algorithm. Historical data is collected from one of NIBE’s active heat pumps running in a field test. The data is processed and experimented on using an algorithm that analyzes how close a certain setting of values for the coefficient used in the filter and sensitivity of the cloud coverage forecast performs compared to an ideal reference. By using an evolutionary algorithm a better solution to the usage of solar energy can be found, compared to the non-evolutionary algorithm, by using a combination of different values for the coefficient in the filter, and also the cloud coverage forecast, which decides when we should change to another value for the filter coefficient.

Generalized hill climbing (GHC) algorithms provide a framework for using local search algorithms to address intractable discrete optimization problems. Many well-known local search algorithms can be formulated as GHC algorithms, including simulated annealing, threshold accepting, Monte Carlo search, and pure local search (among others). This dissertation develops a mathematical framework for simultaneously addressing a set of related discrete optimization problems using GHC algorithms. The resulting algorithms, termed simultaneous generalized hill climbing (SGHC) algorithms, can be applied to a wide variety of sets of related discrete optimization problems. The SGHC algorithm probabilistically moves between these discrete optimization problems according to a problem generation probability function. This dissertation establishes that the problem generation probability function is a stochastic process that satisfies the Markov property. Therefore, given a SGHC algorithm, movement between these discrete optimization problems can be modeled as a Markov chain. Sufficient conditions that guarantee that this Markov chain has a uniform stationary probability distribution are presented. Moreover, sufficient conditions are obtained that guarantee that a SGHC algorithm will visit the globally optimal solution over all the problems in a set of related discrete optimization problems. Computational results are presented with SGHC algorithms for a set of traveling salesman problems. For comparison purposes, GHC algorithms are also applied individually to each traveling salesman problem. These computational results suggest that optimal/near optimal solutions can often be reached more quickly using a SGHC algorithm. / Ph. D.

Generalized hill climbing (GHC) algorithms provide a unifying framework for describing several discrete optimization problem local search heuristics, including simulated annealing and tabu search. A necessary and a sufficient convergence condition for GHC algorithms are presented. The convergence conditions presented in this dissertation are based upon a new iteration classification scheme for GHC algorithms. The convergence theory for particular formulations of GHC algorithms is presented and the implications discussed. Examples are provided to illustrate the relationship between the new convergence conditions and previously existing convergence conditions in the literature. The contributions of the necessary and the sufficient convergence conditions for GHC algorithms are discussed and future research endeavors are suggested. / Ph. D.

Identifying a globally optimal solution for an intractable discrete optimization problem is often cost prohibitive. Therefore, solutions that are within a predetermined threshold are often acceptable in practice. This dissertation introduces the concept of B-acceptable solutions where B is a predetermined threshold for the objective function value. It is difficult to assess a priori the effectiveness of local search algorithms, which makes the process of choosing parameters to improve their performance difficult. This dissertation introduces the B-acceptable solution probability in terms of B-acceptable solutions as a finite-time performance measure for local search algorithms. The B-acceptable solution probability reflects how effectively an algorithm has performed to date and how effectively an algorithm can be expected to perform in the future. The B-acceptable solution probability is also used to obtain necessary asymptotic convergence (with probability one) conditions. Upper and lower bounds for the B-acceptable solution probability are presented. These expressions assume particularly simple forms when applied to specific local search strategies such as Monte Carlo search and threshold accepting. Moreover, these expressions provide guidelines on how to manage the execution of local search algorithm runs. Computational experiments are reported to estimate the probability of reaching a B-acceptable solution for a fixed number of iterations. Logistic regression is applied as a tool to estimate the probability of reaching a B-acceptable solution for values of B close to the objective function value of a globally optimal solution as well as to estimate this objective function value. Computational experiments are reported with logistic regression for pure local search, simulated annealing and threshold accepting applied to instances of the TSP with known optimal solutions. / Ph. D.

Generalized hill climbing (GHC) algorithms are introduced, as a tool to address difficult discrete optimization problems. Particular formulations of GHC algorithms include simulated annealing (SA), local search, and threshold accepting (T A), among. others. A proof of convergence of GHC algorithms is presented, that relaxes the sufficient conditions for the most general proof of convergence for stochastic search algorithms in the literature (Anily and Federgruen [1987]). Proofs of convergence for SA are based on the concept that deteriorating (hill climbing) transitions between neighboring solutions are accepted by comparing a deterministic function of both the solution change cost and a temperature parameter to a uniform (0,1) random variable. GHC algorithms represent a more general model, whereby deteriorating moves are accepted according to a general random variable. Computational results are reported that illustrate relationships that exist between the GHC algorithm's finite-time performance on three problems, and the general random variable formulations used. The dissertation concludes with suggestions for further research. / Ph. D.

Made available in DSpace on 2014-07-29T15:08:20Z (GMT). No. of bitstreams: 1 dissertacao geraldo valeriano eec.pdf: 1215973 bytes, checksum: 4ceb98a1d5250ad33d16a8997882d277 (MD5) Previous issue date: 2009-06-29 / This work presents two methods to solve the problem of Electric Distribution Networks (EDN) with geographical and power restrictions. The high cost of the project involving EDN together with lack of efficient methods when working with real applications justifies the development of this research. Taking into account concepts of heuristic and metaheuristic two methods are proposed: The first is based on the Hill-Climbing (HC) heuristic and the second is based on the Simulated Annealing (SA) metaheuristic. The possible paths are provided by the Delaunay triangulation and it is considered the natural and socio-political obstacles of the site where you want to locate a new energy network. The dimension of the EDN feeders is calculated using the power flow results from the Forward-Backward method. The initial solution is found using an intelligent method. Then the SA metaheuristic and/or HC heuristic are used providing a good solution for a new EDN in comparison with the heuristic used to find the initial solution. A comparison is also made between the two proposed methods / RESUMO Neste trabalho são apresentados dois métodos para resolver o problema de planejamento de rede de distribuição de energia elétrica (RDEE) com restrições geográficas e elétricas. O custo elevado que envolve o projeto de RDEE unido à escassez de métodos eficientes quando se trata de aplicações reais justificam o desenvolvimento desta pesquisa. Considerando os conceitos de heurística e metaheurística são propostos dois métodos: o primeiro é baseado na heurística Hill-Climbing (HC) e o segundo é baseado na metaheurística Simulated Annealing (SA). Os possíveis caminhos são fornecidos pela triangulação de Delaunay e são considerados os obstáculos naturais e políticosociais (restrições geográficas) do local onde se deseja implantar a nova rede de energia elétrica. O dimensionamento dos alimentadores da RDEE é feito utilizando-se do fluxo de potência calculado pelo método Backward-Forward. A solução inicial é encontrada utilizando-se um método inteligente. A metaheurística SA e/ou a heurística HC são então utilizadas, fornecendo uma boa solução para uma nova RDEE, em relação à heurística utilizada para encontrar a solução inicial. Também é realizada uma comparação entre os dois métodos propostos.