Havaalanlarında yer hizmeti veren bir firma için hedef programlama yaklaşımı

Abstract

Çalışmanın amacı; bir çok amaçlı karar verme yöntemi olan Hedef Programlama (HP) yönteminin gerçek hayatta uygulanabilirliğini ve çıkan sonuçların anlamlı olup olmadığını görmektir ve bu nedenle yaşayan bir firma üzerinde HP modeli oluşturulup çözülmüştür. HP yaklaşımının neden kullanıldığı ilgili bölümlerde belirtilmiştir. Yapılan çalışmada, ilk olarak karar verme teorisi anlatılmıştır. Burada, en başta karar vermenin tanımı ve bir problemin karar problemi olabilmesi için gereken şartlardan bahsedilmiştir. Daha sonra, yönetim ve mühendislik açısından karar vermenin öneminden, hem yönetim hem de mühendislik birimlerinin beraber çözmesinin gerektiği problemlerde nasıl davranacakları ve bunun gerekliliği anlatılmıştır. Bu açıklamaları takiben, bölümün esas kısmını oluşturan karar verme sürecinin ve aşamalarının açıklanmasına geçilmiştir. Daha sonra, sürecin aşamaları ayrı ayrı açıklanmış, gerekliliğinden, amacından, nasıl uygulanması gerektiğinden ve birbirleriyle nasıl entegre olacaklarından yani aralarındaki bilgi akışının nasıl olacağından bahsedilmiştir. Bu kısımdan sonra,çok amaçlı karar vermenin günümüz şartlarında gerekliliğine, çok amaçlı karar durumlarının çözümüne ilişkin yöntemlere ve bunların sınıflandırılmasına kısaca değinilmiştir. En son kısımda ise, diğer bölüme kısa bir geçiş niteliğinde olması amacıyla karar analizlerinde HP yaklaşımından bahsedilmiştir. Çalışmanın 3. bölümünde, bir çok amaçlı karar verme yöntemi olan ve yapılmış olan uygulamanın temelim oluşturması nedeniyle HP yaklaşımı anlatılmıştır. İlk olarak HP yönteminin neden kullanıldığından ve üstünlüklerinden bahsedilmiştir. Daha sonra, HP teorisi kısmına geçilmiştir. Bu kısımda, HP tanımı, formülasyonu ve bununla ilgili kavramlar anlatılmıştır. Ayrıca, Doğrusal Hedef Programlama yaklaşımından kısaca bahsedildikten sonra konu ile ilgili iki adet sayısal örnek verilmiştir. Bu kısmı takiben HP' da simpleks yöntem, Doğrusal Olmayan Hedef Programlama ve Etkileşimli Hedef Programlama' dan kısaca bahsedilmiştir. Son kısımda ise HP' nın uygulama alanlarından bahsedilmiştir. Çalışmanın 4. ve amacını oluşturan bölümünde ise HP yaklaşımı kullanılarak bir uygulama yapılmıştır. İlk olarak uygulamanın yapıldığı sektör ve işletme hakkında açıklamalar yapılmıştır. Daha sonra, HP modelinin oluşturulması kısmına geçilmiştir. Burada ilk olarak, modelin amacı, karar vericinin hedefleri ve sektöre ilişkin problemde geçen temel kavramlar açıklanmıştır. Bunu takiben karar değişkenlerinin tanımlanması, hedef ve kısıt denklemlerinin tanımlanması ve katsayılarının hesaplanması anlatılmıştır. Bu işlemler yapıldıktan sonra, model HP formatına uygun bir şekilde formüle edilmiştir. Formüle edilen problem QS (Quant System) adlı programda çözülmüş ve bu çözüme ilişkin tablo halinde verilmiştir. Çıkan sonuçlara göre hedeflenen değerlerden sapma olup olmadığına dair bir analiz yapılarak sonuçların firma bazında değerlendirilmesi kısmına geçilmiştir. Bu kısımda firma bazında, oluşan sonuçların gerçek hayatta nasıl uygulanabileceğine dair yorumlar yapılmıştır. Son olarak ise, problem sonuçları ile gerçek sonuçların karşılaştırılması yapılmıştır.

The purpose of this study is to show implementability of the goal programming which is one of the multiple criteria decision making methods in real world. Therefore, this method has been implemented on a firm which performs in civil aviation sector. Why this method was used has been explained in the study. In the second chapter, decision making theory has been explained in the study. Here by defining the decision as a selection period that direct us to sympathize a certain action, it has been discussed that the decision is dependent on many effects such as;. the psychological status of the decision maker at that moment,. personnel wish and status of the decision maker,. effects of environment,. the possessed facilities,. economical status,. technology and similar facts. Additionally, it has been pointed out that for a problem to be a decision problem;. there should be more than one behavior ways,. the results of each behavior should be different,. there should be some purposes to be realized and the conditions should be realized together. Later, the subject of decision making from the point of management and engineering has been discussed. Here, it has been pointed out that the basic function of management is effective decision making to reach for the purposes and implementation of such, and how great portion of decision making covers the activities of a manager. Also, how the management reaches to the necessary information while making decision has been pointed out and the effectivity of the engineering units has been stressed in this study. Additionally, the importance of the characteristic of the necessary information and how to utilize this information in an effective way in decision making has been pointed out. After such explanations, the fact that the necessity that management and engineering units should work effectively and together and the effect of an effective togetherness to the decision has been pointed out. Also, why the decision analysis should be utilized as a result of all such facts has been explained. vni The stages of decision making period which constitute the main part of the decision making theory have been explained one by one in due order. First of all it has been pointed out that the decision analysis should be on analytical process that utilizes the modern scientific techniques and systematic research methods so as to be assistant to the decision maker in determining the optimum working plan. The stages of process has been listed as follows: 1 ) Definition of The Problem: The necessity that the problem should be defined and the conditions necessary so that the problem becomes a decision problem have been discussed. 2) Design Process of Problem: It consist of the following parts;. Observation of the problem; the importance of being conscious of the symptoms and the realization of this depend on the superiority of the judgment and feeling of the manager and the status of organization of the operation have been pointed out.. Determination of the problem; in this section it is explained that the identity of the problem that has arisen during the observation stage has to be determined with all its details in a final, clear and correct way and how this will be achieved.. Definition of the purposes; it is explained that the decision maker should define his purposes and at which level he wishes to reach to these purposes, in other words, his goals.. Collection and process of data (getting information); the subject of the characteristics that an information should have and how it should be obtained have especially been pointed out. 3) Research and Evaluation of The Options That Suit The Problem: In this section, some information is given in relation to eliminating the unnecessary options by utilizing the information obtained and as the result of comparison of the remaining options among them by means of certain methods and selection of an option that suits the purpose considered. 4) Model Making: Starting from the necessity that the structure and substance of the problem should be abstracted, the importance and the purpose of this stage has been discussed. 5) Selection of The Option That Gives The Optimum Solution: The procedure of selecting the option most convenient to the purpose and while doing these the importance of the criteria to be determined has been explained. 6) Application of The Decisions: In this stage, what the manager should do is explained. 7) Feedback: The importance of having a sound feedback at the stage of application to obtain the requested results has been pointed out. In general, these stages discuss about the necessity and purposes of these stages, how they should be applied and how they will be integrated to each other, in other words, how the information flow will be among them. IX After such explanations, it has been discussed in short that the necessity of multi- objective decision making under the contemporary conditions, the methods related to the solution of multi-objective decision and their classification. Later, for the purpose of a short introduction to the next section, the approach to Goal Programming in decision analysis has been discussed. In the third chapter, goal programming approach has been discussed since this is the most widely used approach in the field of multiple criteria decision making and base of our application. First, treated of definition of goal programming, why goal programming was used and superiority of goal programming. To touch on these subjects briefly; Goal programming is a linear mathematical model in which the optimum attainment of multiple goals is sought within the given decision environment. The decision environment determines the basic components of the model, namely, the decision variables, constraints and the objective function. Solution methods based on numerical criteria are not capable of producing acceptable solutions to problems that involve highly abstract objective criteria such as welfare to the taxpayer, public health, consumer protection and satisfaction, community image of firm and the like. It should be pointed out that numerical solution techniques are still being used for contemporary decision problems by estimating the numerical measures of abstract objective criteria in terms of a convenient numerical value, that's, utilities, profits, costs and so on. However, the process often results in a considerable degree of fabrication and distortion of information in order to express abstract criteria in numerical values. Therefore, the model solution is of very little value o the decision maker. Most real-world decision problems involve multiple conflicting objectives. If we decide to use linear programming to solve decision problems with multiple objectives, we may introduce other objectives (other than the objective function) as model constraints. The linear programming model, however, requires solution must satisfy all constraints. Furthermore, we are assuming that all constraints have equal importance in solving the problems. However, in reality such assumptions are absurd. First of all, it is quite possible that all constraints of the problem cannot be satisfied. Such a problem is called infeasible. Should we abandon a very important management problem because it cannot be solved by linear programming. Next, all constraints do not have equal importance. For example, suppose that the model contains the machine capacity, manpower and union contract constraints. The machine capacity and union contract constraints may be more important and rigid than the manpower constraint. In other words, it is easier to add more employees than to add new heavy machinery or re negotiate with the union. According to all of these reasons, the only alternate method to the numerical approach for problems involving multiple conflicting objective criteria is the ordinal solution approach. Goal programming, based on the ordinal solution approach, appears to be the most appropriate, flexible and powerful technique for complex decision problems involving multiple conflicting objectives. After such explanations, theory, formulation and concepts of goal programming have been studied. Also, after linear goal programming approach has been discussed briefly, given two examples about this subject. In the first example, the computer centre at a university has a plan to purchase personnel computer for one of computer rooms. There are three computer types, that's, three decision variables in example. Based upon the overall priority for each of the alternatives, a goal programming model was formulated. The goal programming model considered the cost constraints (personnel computer hardware, network hardware, software and maintenance and repair over the next three years). The model was used to determine the optimal type and number of computers by satisfying goals given the constraints. In the second example, a profit maximisation of a firm given production capacity and material constraints have been solved by graphical method. The purpose of solving this problem by means of graphical method is to show methodology of the goal programming method. In other part of third chapter, of simplex method of goal programming, non-linear goal programming and interactive goal programming have been explained. In this part; definition of these methods and how they can be used have been discussed. Furthermore, treated of interactive goal programming provides decision makers with more information and allows for more flexibility in considering trade-offs and adjusting goal target levels without appreciably increasing either the computational burden of the procedure or the information processing and evaluation requirements of the decision maker. At the end of third chapter, some of the application areas of goal programming have been presented. Goal programming is widely used approach in real-world as follows: Military application, Water resource planning, Metal cutting planning, Accounting, Academic planning, Portfolio selection, Transportation, Marketing, Product planning and scheduling, Quality control, Project selection, Manpower planning, Assets and liabilities management, To touch remarkable applications; xi Military Application (Aircraft Loading) (15J: The problem is to airlift a specified list of cargo with a minimum number of aircraft loads (chalks) in a prioritised sequence. Goal programming is employed to select from a set of pregenerated feasible loads that subset which yields the minimum number of aircraft loads required to airlift the items in the specified prioritised sequence. In testing with military exercise, the model reduced aircraft loads by 9%, as compared to the manual method presently used and it saved over 240 flying-hour costs. More importantly, the model provides timely planning and improves airlift support of combat operations. Portfolio Selection (Allocation Total Wealth) [18]: In this application, investor has a wealth which he/she wants to allocate among seven assets. Each of the seven possible assets in which the investor can allocate wealth is assigned a decision variable in the linear goal programming model that defines the number of dollars of total wealth that should be allocated in the optimal portfolio. As a result of the linear goal programming model optimised an investor's portfolio. The investor achieved all stated goals. A small experiment was conducted using this problem to examine the potential real- world effectiveness of the proposed linear goal programming model. Two financial advisors from portfolio management firms were selected to prepare their recommendations on the same total wealth allocation problem. The result of the limited comparative study illustrate the possible superiority of the linear goal programming model. At the forth chapter the goal programming approach has been applied to the real world decision problem in the Çelebi Ground Handling Co (CGH) which performs in civil aviation sector. The first explanations are carried out about management and the implementation that is made on sector. Management activity, organisational structure and the status in the sector of the firm have been explained. After that, the brief information about model building and solution processes of the decision problem have been explained. These processes can be arranged as below by their workflow: 1) Definition of objective of the model, 2) Definition of goals of decision maker, 3) Definition of basic concepts, 4) Description decision variables, 5) Definitions of goal and constraint equations and calculation of coefficients, 6) Formulation of model, 7) Solution of the problem, 8) Commendation of results on management firm bases, 9) Comparison of problem results and real results, xu The objective of the model is determined as separately at the period of agreement with the airline firms which are to be considered, namely from CGH point of view which firm has the more added value. CGH' s the largest workstation, which is located in the Atatürk Airport, is considered as the sample for the model. The same application can be considered in the other workstation of CGH as well. The goals of the decision maker is given with the priority sequence: 1) Minimisation of personal expenditure, 2) Maximisation of income, 3) Minimisation of deviation in probable number of flights, Limitations will play the most important part in comparison of goals of themselves. Because all goals and constraints are to be considered together to reach the result. That is, this is the typical problem, which includes multiple conflicting objectives. Some concepts associated with the model that is oriented by the characterisation of the sector are explained at the part of basic concepts. The most important concepts of them are large and narrow body flight concepts. These are divided into categories in accordance with the tonnage of the plain. Explanations related to this discrimination are given as table in related sections. Decision variables are defined as the number of narrow and large body flights on the bases of airline firms. Because these numbers are the main variables in both income and budget calculations. The decision variables which are used for the problem are formed by the number of narrow and large body flights of the Turkish Charter and Tariff firms. After the determination and definition of decision variables, the definition of goals and constraints equations and calculation of their coefficient is carried out goals and constraints equations definition of the problem is sequenced in range as: 1 ) Stationary budget 2) Publication budget 3) Water budget 4) Budget of plane cleaning 5) Budget of cars in use 6) Wireless expenditure budget 7) Sitatex budget 8) Computer stationary budget 9) Contuar budget 10) Incomes 1 1 ) Personal expenditure 12) Yearly expected number of flights belongs to airline firms xm In this section all the coefficients belong to goal and constraint equations which are defined above is explained. From the point of all the definitions and calculations up to now, the problem is formulated as goal programming problem. And definition of all deviation variables (df and d;+). Formulated goal programming problem includes 36 goal and constraint equations and 79 variables. The model has been solved by means of using QS ( Quant System ) software. After 45 iterations, program reached the optimal solution. According to the obtained results the first and the second precedence goals of the decision maker is confirmed. The positive deviation at personal expenditure goal values and the negative deviation income goal values are nullified. Even that negative deviation at personal expenditure, that's, savings. Some deviation occurred in the third precedence goal which is obtained without compensation of these acquired goals. These deviations are the deviation of the number of flights on the bases of the firm. How much deviations on which airline firms at what kind of flights are given in related tables. The meaning of resulting solutions are taken into account on the bases of airline firms. That is, to which airline firms, which type and how many flights is to be serviced by the CGH to reach is goal is explained. Be side that it's commented from the point of CGH and airline firms view that what kind of strategies are to be lead. At the last section, the solution and the real solution of the problem are compared. Here, it is commented on the implementability of the problem result and comparison of the resulting number of the model solution. It is seen from the resulting comparison that the implementability of the results which is obtained by the solution of the model is not the impossible one. Indeed it is so hard to implement all the results because to get into existence of this results needs some struggle and some pre-acceptance. The comments of the experienced persons in sector will be effective on this pre-acceptance. At the results of these explanations, it can be set that; this kind of study is to be great help of determination of strategy to the CGH before the agreements that are performed by CGH. In literature, there's not exist like this kind of study.