In this paper, particle swarm optimization is proposed for finding the global minimum of continuous functions and experimented on benchmark test problems. Particle swarm optimization applied on 21 benchmark test functions, and its solutions are compared to those former proposed approaches: ant colony optimization, a heuristic random optimization, the discrete filled function algorithm, an adaptive random search, dynamic random search technique and random selection walk technique. The implementation of the PSO on several test problems are reported with satisfactory numerical results when compared to previously proposed heuristic techniques. PSO is proved to be successful approach to solve continuous optimization problems.
continuous function optimization; global minimum; heuristic techniques; particle swarm optimization
Bahadır Fatih YILDIRIM, Sultan KUZU, Muhlis ÖZDEMİR
Independently of time and space that provide access to the requested information mobility, internet and mobile devices more widespread usage in daily life increases every day. In this study, the most important instrument mobility on the mobile devices to work compatible hardware and software, allowing controlling and managing mobile operating systems, according to the criteria set by the expert opinion of Multi-Criteria Decision Making (MCDM) approaches are evaluated with VIKOR, in the decision-making process to address uncertainties inherent, the theory of fuzzy systems were included in the decision process decision problem is solved by the method of fuzzy VIKOR. 5 mobile operating systems discussed as decision alternatives have been analyzed by taking expert choice. After all IOS and Android mobile operating systems have taken first place in the set of compromise solution.
Fuzzy VIKOR, Multi-criteria Decision Making, Fuzzy Logic, Mobile Operating Systems, Mobility
Muhlis ÖZDEMİR, Mustafa CAN
Artificial bee colony is one of a swarm intelligence based approach. Artificial bee colony is a metaheuristic method that was inspired by honey bee colonies and based on observing the nourishment behavior of honey bees. Orienteering is a kind of sport which originated in Sweden. Given a set of known locations with starting and ending point, each with a score, a service time, and a time window, a set of vehicle tours that maximizes the total collected scores on condition on turning to the start or end point. In this study artificial bee colony algorithm will be applied to the team orienteering problems with time windows benchmark instances. Artificial bee colony algorithm results for hundred nodes will be compared with iterated local search, variable neighborhood search, fast simulated annealing and slow simulated annealing results in the literature. In addition to these benchmark results new results will be proposed for the fifty nodes problems which might not solved in the literature previously.
Swarm Intelligence, Team Orienteering Problems with Time Windows, Artificial Bee Colony, Meta-Heuristic
Emrah ÖNDER, Bahadır Fatih YILDIRIM, Muhlis ÖZDEMİR
Tourism is the world’s one of fastest growing industry and the largest service sector industry. It is also considered as one of the biggest industry in Turkish economy. Choosing a travel destination is a kind of multi-criteria decision making problem. Relative importance of factors across locations play a crucial role for ranking the destinations. There are several attributes in evaluating competitiveness, including natural resources, transportation, accommodation, blue flagged beaches, cultural resources, reputation, image, popularity, safety, security, health and hygiene, price, quality of cuisine, night life and variety of activities and recreation etc. This study comprised of 13 destination alternatives in four cities (Antalya, Aydın, İzmir, Muğla). These destination alternatives are Alanya, Bodrum, Çeşme, Datça, Didim, Fethiye, Kaş, Kemer, Kumluca, Kuşadası, Marmaris, Manavgat and Serik. Analytic Hierarchy Process (AHP) and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) are multi-criteria decision making (MCDM) methodologies. They have been used extensively for analyzing complex decision problems. These approaches can be used to help decision-makers for prioritizing alternatives and determining the optimal alternative. In analyzing the data, AHP and TOPSIS methodologies are used for the outranking of some of the well known tourism destinations in Turkey. The “safety and security”, “health and hygiene” and “price” are determined as the three most important criteria in the supplier selection process by AHP. Based on TOPSIS analysis the top three of the alternatives in descending order are Alanya, Marmaris and Bodrum. Proposed model results indicate that Alanya is the best alternative with RC value of 0.473.
Multi Criteria Decision Making, Tourism Destination Competitiveness/Ranking, Analytical Hierarchy, Process, TOPSIS
Seyhan NİŞEL, Muhlis ÖZDEMİR
Since the invention of AHP/ANP, several literature review studies have been presented to summarize theoretical developments and different application areas of its techniques. The purpose of this study is to present a comprehensive literature review of AHP and ANP applications in the field of sports. A total of 62 sports related AHP and ANP articles were selected, categorized and analyzed in this study. The findings show that AHP and ANP techniques have successfully been used for performance evaluation of teams, player selection and ranking, team or club performance ranking and coach evaluation in many sport branches.
literature review, AHP, ANP, sport
Çiğdem Arıcıgil ÇİLAN, Nihat TAŞ, Muhlis ÖZDEMİR
Gizli Sınıf Analizi’nde gözlenen tüm değişkenlerin gözlenemeyen gizli bir değişkenin nedeni olduğu kabul edilmektedir. Gizli değişkeni karakterize edebilmek amacıyla gözlenen değişkenler arasındaki ilişkilerin yapıları incelenmektedir. Analizde gözlenen değişkenler arasındaki ilişkinin kaynağı gizli değişkendir. Buna göre gizli değişkenin kontrol değişkeni olarak belirlenmesi durumunda gözlenen değişkenler arasındaki ilişkinin koşullu bağımsız olduğu söylenebilir. Analiz gizli sınıf olasılıkları, koşullu olasılıklar ve üstünlük oranlarının yorumuna dayanır. Bu çalışmada Türkiye İstatistik Kurumu’nun 2012 yılında düzenlediği “Hanehalkı Bilişim Teknolojileri Kullanım Araştırması”’nın mikro verileri temel alınmıştır. Araştırmada öncelikle Türkiye’de internet kullanımının profili tanımsal istatistik ölçülerle belirlenmiş ve Türkiye’de bireylerin internet kullanım faaliyetlerine göre kaç sınıfta toplanabileceği Gizli Sınıf Analizi ile incelenmiştir.
Kategorik Veri Analizi, Gizli Sınıf Analizi, Gizli Sınıf Olasılıkları, Koşullu Olasılıklar
Ergün EROĞLU, Bahadır Fatih YILDIRIM, Muhlis ÖZDEMİR
In Today’s globally competitive environment, business managers are confronted with diverse business problems every day. The human factor is located at the origin of the elements of business efficiency. New employee selection decision directly affects the efficiency of enterprises. The selection process, the process of deciding the expert group who will select employee, evaluating the applied candidates and determining which one of them will be interviewed, according to the identified criterion the proper candidate will be selected. In this study the steps of the ORESTE method being introduced and the ORESTE Multi-criteria decision analysis method which is fewly implemented in Turkish literature, related to various criterion was used for proper personnel selection.
ORESTE, Multi Criteria Decision Making, Personnel Selection
Emrah ÖNDER, Bahadır Fatih YILDIRIM, Muhlis ÖZDEMİR
Combinatorial optimization problems are usually NP-hard and the solution space of them is very large. Therefore the set of feasible solutions cannot be evaluated one by one. Artificial Bee Colony (ABC), Particle Swarm Optimization (PSO) and Genetic Algorithms (GA) are metaheuristic techniques for combinatorial optimization problems. ABC and PSO are swarm intelligence based approaches and they are nature-inspired optimization algorithms. In this study ABC and PSO supported GA techniques were used for finding the shortest route in condition of to visit every city one time but the starting city twice. The problem is a well-known Symmetric Travelling Salesman Problem. Our travelling salesman problem (TSP) consists of 81 cities of Turkey. ABC and PSO-based GA algorithms are applied to solve the travelling salesman problem and results are compared with ant colony optimization (ACO) solution. Our research mainly focused on the application of ABC and PSO based GA algorithms in combinatorial optimization problem. Numerical experiments show that ABC and PSO supported GA are very competitive and have good results compared with the ACO, when it is applied to the regarding problem.