Theses supervised by Prof. Dr. Serap Ulusam Seçkiner
13 theses · Gaziantep University
Diet planning for diabetic patients using mixed integer linear programming
Diabetic diet is an important treatment procedure for diabetic patients. Necessary nutritions should be consumed in proper sizes for a long-life health. In this study a mathematical model has been constructed for diabetic menu planning. Literature about menu planning showed that there is infancy in diabetic menu planning, current studies do not deal with carbohydrate consumption, which is an essential factor for diabetic patients. According to the findings from nutrition treatment recommendations, a mathematical model has been constructed with an objective function that minimizes the total carbohydrate consumption. In this study, it was planned to create a menu with integer programming model and a wide variety of menus were presented to diabetic patients. The model was written separately as 2, 3 and 5-day menus in the GAMS optimization program, and at the last stage, 28-day menus were obtained by solving each of them. This study can be assumed as a preliminary work of the diabetic menu planning with the aim of solving the problem via completing the necessary nutritions parameters.
Agent-based modeling and simulation in hospital energy planning
Healthcare facilities consist of various large buildings that have complex energy systems and high consumption volume of energy, as a result, they emit high carbon emissions. In this direction, the main aim of this thesis is to analyze the energy systems of a hospital in terms of energy converters, energy suppliers, and energy consumption, and consequently, providing flexible solutions that meet energy demands by reducing energy costs and carbon emissions. Initially, to analyze and put forth the pros and cons of the related literature, a systematic literature review was conducted by taking into consideration energy-relevant studies carried out for healthcare facilities. Like our reference hospital, it is difficult that reach short-term consumption data such as hourly load demands due to the absence of energy monitoring systems in particular old buildings. For the purpose of acquiring these short-term datasets, it was developed systematic algorithms that were derived from monthly data with the help of different sources of information, such as the data belong an energy feasibility study conducted in the previous years and various climatic factors. In the third phase, data envelopment analysis-based techniques were applied to measure the energy-based environmental efficiency of the hospital under different input variables on a monthly and yearly basis. In the fourth step, a comprehensive forecasting analysis was carried out with help of different machine learning regression models. In the main phase of the thesis, in the light of the aforementioned analyses, a detailed simulation model of the hospital energy system was improved based on an agent-based modeling approach. The simulation models in which three different scenarios are developed including the current situation are as follows; (1) hospital existing energy infrastructure, (2) adding cogeneration system to the current energy infrastructure, (3) adding cogeneration and photovoltaic systems together to the existing energy infrastructure. Once the running of the simulation models, metaheuristic-based simulation optimization approaches like genetic algorithm was employed to determine optimum cogeneration capacity and optimum number of photovoltaic panels under the aims of minimum system energy cost and carbon emission. Furthermore, a parametric analysis with different test problems was conducted to observe the system reaction to the parameter or condition change. The results indicate that the effective presence of cogeneration and photovoltaic is notable in the decreasing total system cost taking into consideration also lower carbon emissions.
Lean structuring in a sme with model factory
Today, many companies are trying to reduce their costs and establish more efficient systems due to the demands of their customers, such as high quality, low price, speed, and production of personalized products. In this context, learning factories (known as Model Factory in Türkiye), which have examples in different locations and names worldwide, were also established in Türkiye and aimed to support SMEs in this great race. This study aims to create and maintain a lean structure from scratch within the scope of the "Learn-Transform Program" of Model Factories in a company operating in the textile sector and develop company-specific applications. The study's learning model consisting of 22 modules, which includes topics such as 5S, Lean Occupational Health and Safety, and Competence Management, is based on the Gaziantep Model Factory. Based on the presented model, a learning model that forecasts employee retention using artificial neural networks has been developed to reduce the employee turnover rate. A cand company-specific performance evaluation system and company-specific reward systems have been developed. In addition to these processes, it was observed that there was a 25.3% increase in participation in autonomous maintenance processes, a decrease of up to 45% in loss of workforce due to absenteeism, and an improvement of more than 22% in the value stream.
Electrical and electronics goods waste management and optimal warehouse location selection in Turkey
The amount of electronic waste is increasing gradually with the rapid and unconscious consumption of technological products today in Turkey. Within the scope of this study, the current situation of electrical and electronic goods in Turkey has been addressed. The registered collection rate of e-waste formed in our country is low. The main reason for the low collection rates is the unregistered and uncontrolled collection of e-waste. These wastes need to be recovered into the economy. Without a suitable collection structure, it is difficult to recover the resulting electronic waste. In this study, electronic waste (e-waste) that is not collected properly and inconsistently occurred is addressed. It is aimed to establish suitable warehouse locations in the districts close to the organized industrial zones covering certain provinces with developed electronic waste applications and industries in Turkey. There are 28 demand points in the study. To provide the most appropriate service to these demand points, the p-median problem was applied as a location-allocation model. The IBM OPL CPLEX program was used to determine the selection and number of warehouse locations to be opened. As a result of the optimization, suitable warehouse locations to be opened have been decided. It was decided that the optimal number of warehouses was 5. It is difficult for society to leave their wastes from their homes to places where warehouses are planned to be established. It is planned to create waste collection boxes in central places so that all provinces will receive service. The waste collection boxes formed can receive services from the opened warehouse locations.
Dijital hastanelerde verimlilik analizi
Digitalization, introduced to our lives by technology, is pioneering the development of many sectors. The health sector is one of the sectors that has received a great amount of digitalization technology. Today, healthcare service delivery can be provided by different digital applications and digital hospitals. In recent years, the development of digital hospitals has been emphasized in Türkiye as well as on a global scale. Digital hospitals, which are increasing day by day in Türkiye, are graded according to international criteria, and sixth-stage and seventh-stage hospitals are accepted as digital hospitals. However, seventh-stage hospitals have a "zero paper" criterion, unlike sixth-stage hospitals. Accordingly, in this study the efficiency of sixth-stage and seventh-stage hospitals which are selected as samples is analyzed by data envelopment analysis technique under eight different scenarios with identified inputs and outputs based on stationery expenses. In this way, it is measured whether there is a difference in the efficiency of sixth-stage and seventh-stage hospitals due to the "zero paper" criterion. Then, the efficiency scores calculated by data envelopment analysis technique are tested statistically. As a result of data envelopment analysis and statistical analysis, in all scenarios, it has been concluded that the average efficiency of seventh-stage hospitals based on stationery expenses is higher than the average efficiency of sixth-stage hospitals based on stationery expenses. However, since sixth-stage hospitals also have high efficiency, there is no statistically significant difference observed between the hospital-based efficiency scores of sixth-stage and seventh-stage hospitals.
Condition monitoring and control tools for wind energy systems using scada data
Technological developments in wind energy have reduced investment and operating costs. For this reason, wind farms have become more popular around the world. Increasing the share of wind energy in the market has led to the need for easy, inexpensive and effective monitoring and control approaches. In this thesis, various monitoring and control tools are proposed which are cheap and easy to implement in wind farms using existing system data. The first tool is focused on analyzing available data to have a better understand system behavior. Statistical analysis and clustering methods are proposed in this regard. It is necessary to prove that the wind farm operates efficiently and under control. For this reason, the second developed tool is on performance measurements recommended to monitor power generation efficiency. Data Envelopment Analysis, Malmquist Index Approach, and Stochastic Frontier Analysis are proposed to measure power production efficiencies of wind turbines. Forecasting methods have become more important as the wind energy market has increased. Therefore, simple forecasting methods are presented to show the use of available data as the third tool. Also, there is always error in each forecast, so the Particle Filtering approach is suggested to reduce errors. As the last tool, a new training algorithm is developed for multilayer perceptron artificial neural networks which called Antrain ANN. Results are showed that the proposed novel approach can compete with current algorithms.
Problems encountered within the strategic planning in public administration and solution proposals
Strategic Planning approach in public administration in Turkey gained legal status aftermath the promulgation of Public Finance Management and Control Law no.5018 in 2003. The study's aim is to identify the differences and the problems encountered between the first and second strategic planning preparation process and develop solution proposals while analyzing the current situation of the strategic planning system in central public administrations in Turkey. In this study, first of all, the conceptual framework is presented with the literature search. Survey method is used in order to gather datas related to thesis subject. This research primarily hypothesize that ' there is no difference among the institutions related with the preparation, monitoring and evaluation processes of Expired Stratejic Plans (DSESPs) and Current Strategic Plans (YSPs) and the datas gathered from the institutions are evaluated. The results indicate that there is no such a difference among the institutions related with the preparation, monitoring and evaluation processes of Expired Strategic Plan (DSESP) and Current Strategic Plan (YSP). Institutions generally form their strategy documents according to the available minimum conditions inside the strategic planning guide. At the end of the study, proposals such as ensuring effective participation and ownership by senior management have been developed.
Taş ocağı işletmesinde risk analizi ve düzeltici önleyici faaliyetler
Bu tez çalışmasında; inşaatlarda kullanmak üzere taş ve benzeri minerallerin açık ocak madenciliği yöntemiyle çıkarıldığı açık tipteki madenlerin; volkanik kayalar, kaolin, mıcır, granit, kireç taşı, mermer, kum taşı gibi taşların çıkarılması için kullanılan yöntemler ve işletmedeki riskler ele alınmıştır. Bu risklerin önlenmesi ve ortadan kaldırılması için gerekli önleyici ve düzenleyici faaliyetler belirlenmiştir. Bu tez çalışmasında; İş Sağlığı ve Güvenliği ile ilgili genel kavramlar, iş kazaları, meslek hastalıkları, ulusal ve uluslararası kurum ve kuruluşlara ait yasal mevzuatlar ve son olarak Taş Ocağı işleme tesisindeki İş Sağlığı ve Güvenliği açısından genel risk analizi ve buna karşın bu risklerin ortadan kaldırılmasına yönelik önlemlerin nasıl alınacağına dair önlem ve öneriler geniş olarak ele alınmıştır.
Optimum determination of number, capacity and location of bioenergy power plant/s by using geographic information system: A case study in Bismil
Bismil district of Diyarbakir is a region where agricultural activities are very common. About 60% of the terrain is irrigable. Cotton and corn are mostly grown in these irrigable lands. In this study, a mixed integer linear programming (MILP) model and a metaheuristic algorithm which optimize the numbers, locations and capacities of the CCHP (tri-generation=combined cooling, heating and power) facility(ies) that produces energy from cotton and corn residues were developed and solved. Both technics try to maximize the net present value of the potential investment by comparing the trade-offs between the agricultural residues/endproduct transportation costs, facility(ies) and pipeline installation costs. All points (supply, candidate power plant, heat demand, electricity distribution) were determined in Geographic Information System. However, size of the problem was reduced to test the viability of the proposed MILP model. After proving viability of the MILP model, a metaheuristic algorithm (Simulated Annealing) was employed for a real-world problem in Bismil. As annual revenue, electricity, the heating/cooling energy and the fertilizer (biochar) sale; as for annual expenses, transportation of raw materials and heating/cooling energy, losses during heat and electricity transportation, operational maintenance costs of facilities and pipelines were considered. Initial investment cost includes facility, pipeline and electricity transmission installation costs. The model proposed 1 facility with an installed power of about 29 MW. Cooling energy demand of 41 settlements, both cooling and heat energy demand of 9 settlements were met. The net present value of the investment is almost 260 million euro and the repayment period is 1,3 years.
Technology transfer and development with public procurement SSO smart purchasing model example
It is vital that our country believe in primitive unmanned sales as leverage. But there are some problems in front of it. Through public procurement; and an entrepreneurship ecosystem and public procurement research for innovation, localization, technology transfer and entrepreneurship, a business model innovation in the Canvas Business Model was proposed and SSO Techno Catalog Platform was prepared as solution for the first stage solution, solved by means of the State Supply Office (SSO) of the Central Purchasing Agency of Turkey. SSO Techno Catalog Platform is a platform that enhances the marketing competencies of our entrepreneurs and accelerates the process of commercialization and brand processes. The SSO Techno Catalog Platform is a start-up training for managing the public procurement in our country with the understanding of technology management, for the Technology Readiness Level and the entrepreneurial ecosystem in a common language. Key Words: Public Procurement, SSO Techno Catalog Platform, Canvas Business Model, Commercialization, Technology Transfer
Developing and implementing a wage assessment method based on the 5S performance score: an application of steel construction factory
In today's competitive environment, the trend is to respond quickly to customer demand and deliver high-quality, low-priced products. To be a global company, it must be the goal of all managers to remove wasted resources and achieve an effective workflow. The most important condition of providing an efficient environment is the creation of a clean and tidy environment that is free of possible hazards. Our working environments tend to be irregular and untidy at all times, due to the dynamic nature of their existence. One of the basic tools of providing this arrangement and order in the workplace is the 5S applications. In this context, especially 5S application which is being used in the industry has some superiorities such as easily carrying out, increasing the participation and satisfaction of the employees. In job evaluation studies, jobs are evaluated according to predefined factors and then their values are defined with respect to each other. Then wage curves are drawn and wage brackets are set for each point interval over these values. In this study, 5S scores were used in the evaluation of success (merit) factor in the internal structure of wage brackets in order to increase the employee's belief in the objectivity of the wage assessment system. Thus, a new wage assessment method, in which both lean manufacturing and job analysis and evaluation system are integrated, has been developed and implemented. This application gave solutions to chronic problems of the applied factory, where workers were slowing or stopping works because of wage increment dissatisfactions in additionally to unhealthy, dirty and untidy work environment and low occupational safety preventions.
Empirical modelling, control and optimization of lubricant recycling unit
This study consists of two phases. In the first phase, four different forecasting approaches were applied to the chemical processes found in the lubrican recycling unit. The predictive methods used are Multiple Linear Regression, Multiple Nonlinear Regression, Artificial Neural Networks and Ridge Regression. The input variables of models are temperature and pressure, and the output variables of the unit are quantity and quality. Both quantities and qualities of useful end products are classified as light gas oil, medium gas oil, and heavy gas oil and are kept in the daily record of the chemical unit. Four different forecasting models have been compared in term of their performance. The artificial neural network approach showed best predictive model with the least error rate. In the second phase of the study, the non-linear mathematical optimization model generated the optimal temperature and pressure values in order to maximize the quantity and quality of light, medium, and heavy gas oils. Thus, optimum parameters have been obtained to help eliminate at least 4300 litres of wasted oil per week and production has been increased from 57% to over 70%. As a result, it has been observed that controlling the chemical processes by nonlinear mathematical modelling increases the production quantity. Key Words: Empirical modelling, Multiple Linear Regression, Multiple Nonlinear Regression, Artificial Neural Networks, Ridge Regression.
A health-based metaheuristic algorithm: Cholesterol algorithm
Optimization is as old as the history of the universe. It can be seen anywhere from the farthest point of the universe to the atomic particle. Optimization can sometimes be seen in an engineering design and sometimes in an ant's search for food. The main purpose of all is to find the best possible result under certain conditions. The universe, nature, animals solve their problems easily by instinct. For this reason, people try to solve complex problems by imitating the universe and nature. They developed metaheuristics to solve complex optimization problems. There are many algorithms in the literature that are inspired by animals and nature. However, the number of algorithms inspired by the human body such as artificial neural networks and artificial immune systems are very few. In this thesis, a new health-based metaheuristic algorithm has been developed, inspired by the cholesterol metabolism in the human body. The cholesterol algorithm is modeled using the cholesterol levels in the human body. This thesis focuses on the performance of the cholesterol algorithm on unconstrained continuous optimization problems. By keeping the structure of the cholesterol algorithm simple, a faster and more flexible algorithm has been developed. For the performance analysis of the cholesterol algorithm, 23 comparison tests were used and the test results were compared with 11 different algorithms.