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.

Automated, adaptive methods for forest inventory

Forest inventories are used to estimate forest characteristics and the condition of forest for many different applications: operational tree logging for forest industry, forest health state estimation, carbon balance estimation, land-cover and land use analysis in order to avoid forest degradation etc. Recent inventory methods are strongly based on remote sensing data combined with field sample measurements, which are used to define estimates covering the whole area of interest. Remote sensing data from satellites, aerial photographs or aerial laser scannings are used, depending on the scale of inventory. To be applicable in operational use, forest inventory methods need to be easily adjusted to local conditions of the study area at hand. All the data handling and parameter tuning should be objective and automated as much as possible. The methods also need to be robust when applied to different forest types. Since there generally are no extensive direct physical models connecting the remote sensing data from different sources to the forest parameters that are estimated, mathematical estimation models are of "black-box" type, connecting the independent auxiliary data to dependent response data with linear or nonlinear arbitrary models. To avoid redundant complexity and over-fitting of the model, which is based on up to hundreds of possibly collinear variables extracted from the auxiliary data, variable selection is needed. To connect the auxiliary data to the inventory parameters that are estimated, field work must be performed. In larger study areas with dense forests, field work is expensive, and should therefore be minimized. To get cost-efficient inventories, field work could partly be replaced with information from formerly measured sites, databases. The work in this thesis is devoted to the development of automated, adaptive computation methods for aerial forest inventory. The mathematical model parameter definition steps are automated, and the cost-efficiency is improved by setting up a procedure that utilizes databases in the estimation of new area characteristics.

Developing of the Variable Speed Drive -control in simulated parallel pumping systems

This thesis presents briefly the basic operation and use of centrifugal pumps and parallel pumping applications. The characteristics of parallel pumping applications are compared to circuitry, in order to search analogy between these technical fields. The purpose of studying circuitry is to find out if common software tools for solving circuit performance could be used to observe parallel pumping applications. The empirical part of the thesis introduces a simulation environment for parallel pumping systems, which is based on circuit components of Matlab Simulink —software. The created simulation environment ensures the observation of variable speed controlled parallel pumping systems in case of different controlling methods. The introduced simulation environment was evaluated by building a simulation model for actual parallel pumping system at Lappeenranta University of Technology. The simulated performance of the parallel pumps was compared to measured values of the actual system. The gathered information shows, that if the initial data of the system and pump perfonnance is adequate, the circuitry based simulation environment can be exploited to observe parallel pumping systems. The introduced simulation environment can represent the actual operation of parallel pumps in reasonably accuracy. There by the circuitry based simulation can be used as a researching tool to develop new controlling ways for parallel pumps.

In search of information systems value : a case study of the EHR benefits realisation efforts of three Swedish hospitals.

The Swedish public health care organisation could very well be undergoing its most significant change since its specialisation during the late 19th and early 20th century. At the heart of this change is a move from using manual patient journals to electronic health records (EHR). EHR are complex integrated organisational wide information systems (IS) that promise great benefits and value as well as presenting great challenges to the organisation. The Swedish public health care is not the first organisation to implement integrated IS, and by no means alone in their quest for realising the potential benefits and value that it has to offer. As organisations invest in IS they embark on a journey of value-creation and capture. A journey where a costbased approach towards their IS-investments is replaced with a value-centric focus, and where the main challenges lie in the practical day-to-day task of finding ways to intertwine technology, people and business processes. This has however proven to be a problematic task. The problematic situation arises from a shift of perspective regarding how to manage IS in order to gain value. This is a shift from technology delivery to benefits delivery; from an ISimplementation plan to a change management plan. The shift gives rise to challenges related to the inability of IS and the elusiveness of value. As a response to these challenges the field of IS-benefits management has emerged offering a framework and a process in order to better understand and formalise benefits realisation activities. In this thesis the benefits realisation efforts of three Swedish hospitals within the same county council are studied. The thesis focuses on the participants of benefits analysis projects; their perceptions, judgments, negotiations and descriptions of potential benefits. The purpose is to address the process where organisations seek to identify which potential IS-benefits to pursue and realise, this in order to better understand what affects the process, so that realisation actions of potential IS-benefits could be supported. A qualitative case study research design is adopted and provides a framework for sample selection, data collection, and data analysis. It also provides a framework for discussions of validity, reliability and generalizability. Findings displayed a benefits fluctuation, which showed that participants’ perception of what constituted potential benefits and value changed throughout the formal benefits management process. Issues like structure, knowledge, expectation and experience affected perception differently, and this in the end changed the amount and composition of potential benefits and value. Five dimensions of benefits judgment were identified and used by participants when finding accommodations of potential benefits and value to pursue. Identified dimensions affected participants’ perceptions, which in turn affected the amount and composition of potential benefits. During the formal benefits management process participants shifted between judgment dimensions. These movements emerged through debates and interactions between participants. Judgments based on what was perceived as expected due to one’s role and perceived best for the organisation as a whole were the two dominant benefits judgment dimensions. A benefits negotiation was identified. Negotiations were divided into two main categories, rational and irrational, depending on participants’ drive when initiating and participating in negotiations. In each category three different types of negotiations were identified having different characteristics and generating different outcomes. There was also a benefits negotiation process identified that displayed management challenges corresponding to its five phases. A discrepancy was also found between how IS-benefits are spoken of and how actions of IS benefits realisation are understood. This was a discrepancy between an evaluation and a realisation focus towards IS value creation. An evaluation focus described IS-benefits as well-defined and measurable effects and a realisation focus spoke of establishing and managing an on-going place of value creation. The notion of valuescape was introduced in order to describe and support the understanding of IS value creation. Valuescape corresponded to a realisation focus and outlined a value configuration consisting of activities, logic, structure, drivers and role of IS.

Nested Rollout Policy Adaptation for optimizing vehicle selection in complex VRPs

The goal of Vehicle Routing Problems (VRP) and their variations is to transport a set of orders with the minimum number of vehicles at least cost. Most approaches are designed to solve specific problem variations independently, whereas in real world applications, different constraints are handled concurrently. This research extends solutions obtained for the traveling salesman problem with time windows to a much wider class of route planning problems in logistics. The work describes a novel approach that:  supports a heterogeneous fleet of vehicles  dynamically reduces the number of vehicles  respects individual capacity restrictions  satisfies pickup and delivery constraints  takes Hamiltonian paths (rather than cycles) The proposed approach uses Monte-Carlo Tree Search and in particular Nested Rollout Policy Adaptation. For the evaluation of the work, real data from the industry was obtained and tested and the results are reported.