Accuracy improvement of a demand forecasting model

This research concerns different statistical methods that assist to increase the demand forecasting accuracy of company X’s forecasting model. Current forecasting process was analyzed in details. As a result, graphical scheme of logical algorithm was developed. Based on the analysis of the algorithm and forecasting errors, all the potential directions for model future improvements in context of its accuracy were gathered into the complete list. Three improvement directions were chosen for further practical research, on their basis, three test models were created and verified. Novelty of this work lies in the methodological approach of the original analysis of the model, which identified its critical points, as well as the uniqueness of the developed test models. Results of the study formed the basis of the grant of the Government of St. Petersburg.

Evaluating and improving demand forecast practices for B2B and B2C sectors

Demand forecasting is one of the fundamental managerial tasks. Most companies do not know their future demands, so they have to make plans based on demand forecasts. The literature offers many methods and approaches for producing forecasts. Former literature points out that even though many forecasting methods and approaches are available, selecting a suitable approach and implementing and managing it is a complex cross-functional matter. However, it’s relatively rare that researches are focused on the differences in forecasting between consumer and industrial companies. The aim of this thesis is to investigate the potential of improving demand forecasting practices for B2B and B2C sectors in the global supply chains. Business to business (B2B) sector produces products for other manufacturing companies. On the other hand, consumer (B2C) sector provides goods for individual buyers. Usually industrial sector have a lower number of customers and closer relationships with them. The research questions of this thesis are: 1) What are the main differences and similarities in demand planning between B2B and B2C sectors? 2) How the forecast performance for industrial and consumer companies can be improved? The main methodological approach in this study is design science, where the main objective is to develop tentative solutions to real-life problems. The research data has been collected from a case company. Evaluation and improving in organizing demand forecasting can be found in three interlinked areas: 1) demand planning operational environment, 2) demand forecasting techniques, 3) demand information sharing scenarios. In this research current B2B and B2C demand practices are presented with further comparison between those two sectors. It was found that B2B and B2C sectors have significant differences in demand practices. This research partly filled the theoretical gap in understanding the difference in forecasting in consumer and industrial sectors. In all these areas, examples of managerial problems are described, and approaches for mitigating these problems are outlined.

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.

Households’ willingness to engage in demand response in the Finnish retail electricity market: an empirical study

If electricity users adjusted their consumption patterns according to time-variable electricity prices or other signals about the state of the power system, generation and network assets could be used more efficiently, and matching intermittent renewable power generation with electricity demand would be facilitated. This kind of adjustment of electricity consumption, or demand response, may be based on consumers’ decisions to shift or reduce electricity use in response to time-variable electricity prices or on the remote control of consumers’ electric appliances. However, while demand response is suggested as a solution to many issues in power systems, actual experiences from demand response programs with residential customers are mainly limited to short pilots with a small number of voluntary participants, and information about what kinds of changes consumers are willing and able to make and what motivates these changes is scarce. This doctoral dissertation contributes to the knowledge about what kinds of factors impact on residential consumers’ willingness and ability to take part in demand response. Saving opportunities calculated with actual price data from the Finnish retail electricity market are compared with the occurred supplier switching to generate a first estimate about how large savings could trigger action also in the case of demand response. Residential consumers’ motives to participate in demand response are also studied by a web-based survey with 2103 responses. Further, experiences of households with electricity consumption monitoring systems are discussed to increase knowledge about consumers’ interest in getting more information on their electricity use and adjusting their behavior based on it. Impacts of information on willingness to participate in demand response programs are also approached by a survey for experts of their willingness to engage in demand response activities. Residential customers seem ready to allow remote control of electric appliances that does not require changes in their everyday routines. Based on residents’ own activity, the electricity consuming activities that are considered shiftable are very limited. In both cases, the savings in electricity costs required to allow remote control or to engage in demand response activities are relatively high. Nonmonetary incentives appeal to fewer households.

Heat Demand Forecasting Models’ Development: Use of Data Mining Tools in SQL Server Analysis Services

This thesis introduces heat demand forecasting models which are generated by using data mining algorithms. The forecast spans one full day and this forecast can be used in regulating heat consumption of buildings. For training the data mining models, two years of heat consumption data from a case building and weather measurement data from Finnish Meteorological Institute are used. The thesis utilizes Microsoft SQL Server Analysis Services data mining tools in generating the data mining models and CRISP-DM process framework to implement the research. Results show that the built models can predict heat demand at best with mean average percentage errors of 3.8% for 24-h profile and 5.9% for full day. A deployment model for integrating the generated data mining models into an existing building energy management system is also discussed.

Modelling the exothermic heat and the demand of cooling in paper and pulp industry biological wastewater treatment

This master thesis presents a study on the requisite cooling of an activated sludge process in paper and pulp industry. The energy consumption of paper and pulp industry and it’s wastewater treatment plant in particular is relatively high. It is therefore useful to understand the wastewater treatment process of such industries. The activated sludge process is a biological mechanism which degrades carbonaceous compounds that are present in waste. The modified activated sludge model constructed here aims to imitate the bio-kinetics of an activated sludge process. However, due to the complicated non-linear behavior of the biological process, modelling this system is laborious and intriguing. We attempt to find a system solution first using steady-state modelling of Activated Sludge Model number 1 (ASM1), approached by Euler’s method and an ordinary differential equation solver. Furthermore, an enthalpy study of paper and pulp industry’s vital pollutants was carried out and applied to revise the temperature shift over a period of time to formulate the operation of cooling water. This finding will lead to a forecast of the plant process execution in a cost-effective manner and management of effluent efficiency. The final stage of the thesis was achieved by optimizing the steady state of ASM1.

Offshoring and outsourcing under demand uncertainty -a supply chain management analysis

Outsourcing and offshoring or any combinations of these have not just become a popular phenomenon, but are viewed as one of the most important management strategies due to the new possibilities from globalization. They have been seen as a possibility to save costs and improve customer service. Executing offshoring and offshore outsourcing successfully can be more complex than initially expected. Potential cost savings resulting from of offshoring and offshore outsourcing are often based on lower manufacturing costs. However, these benefits might be conflicted by a more complex supply chain with service level challenges that can respectively increase costs. Therefore analyzing the total cost effects of offshoring and outsourcing is necessary. The aim of this Master´s Thesis was to to construct a total cost model using academic literature to calculate the total costs and analyze the reasonability of offshoring and offshore outsourcing production of a case company compared to insourcing production. The research data was mainly quantitative and collected mainly from the case company past sales and production records. In addition management level interviews from the case company were conducted. The information from these interviews was used for the qualification of the necessary quantitative data and adding supportive information that could not be gathered from the quantitative data. Both data collection and analysis were guided by a theoretical frame of reference that was based on academic literature concerning offshoring and outsourcing, statistical calculation of demand and total costs. The results confirm the theories that offshoring and offshore outsourcing would reduce total costs as both offshoring and offshore outsourcing options result in lower total annual costs than insourcing mainly due to lower manufacturing costs. However, increased demand uncertainty would make the alternative of offshore outsourcing more risky and difficult to manage. Therefore when assessing the overall impact of the alternatives, offshoring is the most preferable option. As the main cost savings in offshore outsourcing came from lower manufacturing costs, more specifically labour costs, the logistics costs in this case company did not have an essential effect in total costs. The management should therefore pay attention initially to manufacturing costs and then logistics costs when choosing the best production sourcing option for the company.