Context-specific independence in graphical models

The theme of this thesis is context-speci c independence in graphical models. Considering a system of stochastic variables it is often the case that the variables are dependent of each other. This can, for instance, be seen by measuring the covariance between a pair of variables. Using graphical models, it is possible to visualize the dependence structure found in a set of stochastic variables. Using ordinary graphical models, such as Markov networks, Bayesian networks, and Gaussian graphical models, the type of dependencies that can be modeled is limited to marginal and conditional (in)dependencies. The models introduced in this thesis enable the graphical representation of context-speci c independencies, i.e. conditional independencies that hold only in a subset of the outcome space of the conditioning variables. In the articles included in this thesis, we introduce several types of graphical models that can represent context-speci c independencies. Models for both discrete variables and continuous variables are considered. A wide range of properties are examined for the introduced models, including identi ability, robustness, scoring, and optimization. In one article, a predictive classi er which utilizes context-speci c independence models is introduced. This classi er clearly demonstrates the potential bene ts of the introduced models. The purpose of the material included in the thesis prior to the articles is to provide the basic theory needed to understand the articles.

Production planning and optimization of a pulp mill: introduction of the optimisation based scheduling software for pulp industry

The objective of this project was to introduce a new software product to pulp industry, a new market for case company. An optimization based scheduling tool has been developed to allow pulp operations to better control their production processes and improve both production efficiency and stability. Both the work here and earlier research indicates that there is a potential for savings around 1-5%. All the supporting data is available today coming from distributed control systems, data historians and other existing sources. The pulp mill model together with the scheduler, allows what-if analyses of the impacts and timely feasibility of various external actions such as planned maintenance of any particular mill operation. The visibility gained from the model proves also to be a real benefit. The aim is to satisfy demand and gain extra profit, while achieving the required customer service level. Research effort has been put both in understanding the minimum features needed to satisfy the scheduling requirements in the industry and the overall existence of the market. A qualitative study was constructed to both identify competitive situation and the requirements vs. gaps on the market. It becomes clear that there is no such system on the marketplace today and also that there is room to improve target market overall process efficiency through such planning tool. This thesis also provides better overall understanding of the different processes in this particular industry for the case company.

Efficient Optimization Algorithms for Nonlinear Data Analysis

Identification of low-dimensional structures and main sources of variation from multivariate data are fundamental tasks in data analysis. Many methods aimed at these tasks involve solution of an optimization problem. Thus, the objective of this thesis is to develop computationally efficient and theoretically justified methods for solving such problems. Most of the thesis is based on a statistical model, where ridges of the density estimated from the data are considered as relevant features. Finding ridges, that are generalized maxima, necessitates development of advanced optimization methods. An efficient and convergent trust region Newton method for projecting a point onto a ridge of the underlying density is developed for this purpose. The method is utilized in a differential equation-based approach for tracing ridges and computing projection coordinates along them. The density estimation is done nonparametrically by using Gaussian kernels. This allows application of ridge-based methods with only mild assumptions on the underlying structure of the data. The statistical model and the ridge finding methods are adapted to two different applications. The first one is extraction of curvilinear structures from noisy data mixed with background clutter. The second one is a novel nonlinear generalization of principal component analysis (PCA) and its extension to time series data. The methods have a wide range of potential applications, where most of the earlier approaches are inadequate. Examples include identification of faults from seismic data and identification of filaments from cosmological data. Applicability of the nonlinear PCA to climate analysis and reconstruction of periodic patterns from noisy time series data are also demonstrated. Other contributions of the thesis include development of an efficient semidefinite optimization method for embedding graphs into the Euclidean space. The method produces structure-preserving embeddings that maximize interpoint distances. It is primarily developed for dimensionality reduction, but has also potential applications in graph theory and various areas of physics, chemistry and engineering. Asymptotic behaviour of ridges and maxima of Gaussian kernel densities is also investigated when the kernel bandwidth approaches infinity. The results are applied to the nonlinear PCA and to finding significant maxima of such densities, which is a typical problem in visual object tracking.

Distribution network design and optimization

The objective of this thesis is to examine distribution network designs and modeling practices and create a framework to identify best possible distribution network structure for the case company. The main research question therefore is: How to optimize case company’s distribution network in terms of customer needs and costs? Theory chapters introduce the basic building blocks of the distribution network design and needed calculation methods and models. Framework for the distribution network projects was created based on the theory and the case study was carried out by following the defined framework. Distribution network calculations were based on the company’s sales plan for the years 2014 - 2020. Main conclusions and recommendations were that the new Asian business strategy requires high investments in logistics and the first step is to open new satellite DC in China as soon as possible to support sales and second possible step is to open regional DC in Asia within 2 - 4 years.

Stability Analysis in Multicriteria Discrete Portfolio Optimization.

Almost every problem of design, planning and management in the technical and organizational systems has several conflicting goals or interests. Nowadays, multicriteria decision models represent a rapidly developing area of operation research. While solving practical optimization problems, it is necessary to take into account various kinds of uncertainty due to lack of data, inadequacy of mathematical models to real-time processes, calculation errors, etc. In practice, this uncertainty usually leads to undesirable outcomes where the solutions are very sensitive to any changes in the input parameters. An example is the investment managing. Stability analysis of multicriteria discrete optimization problems investigates how the found solutions behave in response to changes in the initial data (input parameters). This thesis is devoted to the stability analysis in the problem of selecting investment project portfolios, which are optimized by considering different types of risk and efficiency of the investment projects. The stability analysis is carried out in two approaches: qualitative and quantitative. The qualitative approach describes the behavior of solutions in conditions with small perturbations in the initial data. The stability of solutions is defined in terms of existence a neighborhood in the initial data space. Any perturbed problem from this neighborhood has stability with respect to the set of efficient solutions of the initial problem. The other approach in the stability analysis studies quantitative measures such as stability radius. This approach gives information about the limits of perturbations in the input parameters, which do not lead to changes in the set of efficient solutions. In present thesis several results were obtained including attainable bounds for the stability radii of Pareto optimal and lexicographically optimal portfolios of the investment problem with Savage's, Wald's criteria and criteria of extreme optimism. In addition, special classes of the problem when the stability radii are expressed by the formulae were indicated. Investigations were completed using different combinations of Chebyshev's, Manhattan and Hölder's metrics, which allowed monitoring input parameters perturbations differently.

Optimization and Measuring Techniques for Collect-and-Place Machines in Printed Circuit Board Industry

This thesis considers optimization problems arising in printed circuit board assembly. Especially, the case in which the electronic components of a single circuit board are placed using a single placement machine is studied. Although there is a large number of different placement machines, the use of collect-and-place -type gantry machines is discussed because of their flexibility and increasing popularity in the industry. Instead of solving the entire control optimization problem of a collect-andplace machine with a single application, the problem is divided into multiple subproblems because of its hard combinatorial nature. This dividing technique is called hierarchical decomposition. All the subproblems of the one PCB - one machine -context are described, classified and reviewed. The derived subproblems are then either solved with exact methods or new heuristic algorithms are developed and applied. The exact methods include, for example, a greedy algorithm and a solution based on dynamic programming. Some of the proposed heuristics contain constructive parts while others utilize local search or are based on frequency calculations. For the heuristics, it is made sure with comprehensive experimental tests that they are applicable and feasible. A number of quality functions will be proposed for evaluation and applied to the subproblems. In the experimental tests, artificially generated data from Markov-models and data from real-world PCB production are used. The thesis consists of an introduction and of five publications where the developed and used solution methods are described in their full detail. For all the problems stated in this thesis, the methods proposed are efficient enough to be used in the PCB assembly production in practice and are readily applicable in the PCB manufacturing industry.

Differential evolution based classification with pool of distances and aggregation operators

The objective of this thesis is to develop and generalize further the differential evolution based data classification method. For many years, evolutionary algorithms have been successfully applied to many classification tasks. Evolution algorithms are population based, stochastic search algorithms that mimic natural selection and genetics. Differential evolution is an evolutionary algorithm that has gained popularity because of its simplicity and good observed performance. In this thesis a differential evolution classifier with pool of distances is proposed, demonstrated and initially evaluated. The differential evolution classifier is a nearest prototype vector based classifier that applies a global optimization algorithm, differential evolution, to determine the optimal values for all free parameters of the classifier model during the training phase of the classifier. The differential evolution classifier applies the individually optimized distance measure for each new data set to be classified is generalized to cover a pool of distances. Instead of optimizing a single distance measure for the given data set, the selection of the optimal distance measure from a predefined pool of alternative measures is attempted systematically and automatically. Furthermore, instead of only selecting the optimal distance measure from a set of alternatives, an attempt is made to optimize the values of the possible control parameters related with the selected distance measure. Specifically, a pool of alternative distance measures is first created and then the differential evolution algorithm is applied to select the optimal distance measure that yields the highest classification accuracy with the current data. After determining the optimal distance measures for the given data set together with their optimal parameters, all determined distance measures are aggregated to form a single total distance measure. The total distance measure is applied to the final classification decisions. The actual classification process is still based on the nearest prototype vector principle; a sample belongs to the class represented by the nearest prototype vector when measured with the optimized total distance measure. During the training process the differential evolution algorithm determines the optimal class vectors, selects optimal distance metrics, and determines the optimal values for the free parameters of each selected distance measure. The results obtained with the above method confirm that the choice of distance measure is one of the most crucial factors for obtaining higher classification accuracy. The results also demonstrate that it is possible to build a classifier that is able to select the optimal distance measure for the given data set automatically and systematically. After finding optimal distance measures together with optimal parameters from the particular distance measure results are then aggregated to form a total distance, which will be used to form the deviation between the class vectors and samples and thus classify the samples. This thesis also discusses two types of aggregation operators, namely, ordered weighted averaging (OWA) based multi-distances and generalized ordered weighted averaging (GOWA). These aggregation operators were applied in this work to the aggregation of the normalized distance values. The results demonstrate that a proper combination of aggregation operator and weight generation scheme play an important role in obtaining good classification accuracy. The main outcomes of the work are the six new generalized versions of previous method called differential evolution classifier. All these DE classifier demonstrated good results in the classification tasks.

Optimization of a dynamic district heating power plant

The purpose of this Thesis is to find the most optimal heat recovery solution for Wärtsilä’s dynamic district heating power plant considering Germany energy markets as in Germany government pays subsidies for CHP plants in order to increase its share of domestic power production to 25 % by 2020. Different heat recovery connections have been simulated dozens to be able to determine the most efficient heat recovery connections. The purpose is also to study feasibility of different heat recovery connections in the dynamic district heating power plant in the Germany markets thus taking into consideration the day ahead electricity prices, district heating network temperatures and CHP subsidies accordingly. The auxiliary cooling, dynamical operation and cost efficiency of the power plant is also investigated.

Logistics Cost Optimization of Road Freight Shipments in a Consumer Goods Company

In the industry of the case company, transportation and warehousing costs account for more than 10% of the total cost which is more than on average. A Finnish company has an understanding that by sending larger shipments in parcels, they could save tens of thousands of euros annually in freight costs in Finland’s domestic shipments. To achieve these savings and optimize total logistics cost, company’s interest is to find out which is the cost efficient way of shipping road shipments of certain volumes; in parcel boxes or on pallets, and what should be the split volume determining the shipment type. Distribution center (DC) costs affect this decision and therefore they need to be also evaluated to determine the total logistics cost savings. Main results were achieved by executing activity-based costing-calculations including DC and road freight costs to determine the ideal split volume with which the total logistics cost is optimal. Calculations were done for Finland’s DC, separately for two main road freight destinations, Finland and Sweden, which cover 50% of road shipment spend. Data for calculations was collected both manually and automatically from various internal and external sources, such as the company ERP system and logistics service providers’ (LSP) reporting. DC processes were studied in practice and compared to model processes. Currently used freight rates were compared to existing pricing models and freight service tendering process was evaluated by participating in the process and comparing it to the models based on literature. The results show that the potential savings are not as significant as the company hoped for, mainly because of packing work increasing DC labor cost. Annual savings by setting ideal split volume per country would account for 0,4 % of the warehousing and transportation costs of shipments in scope of this thesis. Split volume should be set separately for each route, mainly because the pricing model for road freight is different in each country. For some routes bigger parcels should be sent but for some routes pallets should be used more. Next step is to do these calculations for remaining routes to determine total savings potential. Other findings show that the processes in the DC are designed well and the company could achieve savings by executing tenders more efficiently. Company should also pay more attention to parcel pricing and packing the shipments accordingly.