PROPOSAL OF SPECIFIC MEASURES IN THE CONTEXT OF STREAMLINING OPERATING COSTS IN A SELECTED COMPANY

The aim of this research is to propose particular measures aimed at streamlining costs in terms of operations in a selected company. A specific proposal is based on the concept related to using services of the transport databank. As at least two out of five major carriers of the investigated company resell shipments to other carriers on such portals, the company could reduce the transport costs by entering shipments into the transport databank on its own. One out of three selected providers will be chosen using multi-criteria evaluation methods, specifically, the Analytic Hierarchy Process (hereinafter referred to as AHP) and the Base-criterion method. Determination of the weights of criteria will be carried out using the Saaty method of quantitative pair wise comparison and the Fuller pair wise comparison method. In this paper, the presented methods are applied for the example of a specific company specialized in the manufacture of metal storage racks and steel structures. In the Czech Republic, the company employs 65 employees in production, storage and administration. Based on the analytical evaluation of the current situation of the company, relevant measures will be proposed with a required effect on the effectiveness of such an enterprise.


INTRODUCTION
One of the main financial objectives of a company is to achieve the highest profits possible. The sub-objective is usually to make efforts to maximize their market value. These objectives can be achieved by proper management of business costs. Currently, the market is characterized by strong competition, especially in terms of small-and medium-sized enterprises, which are particularly forced to reduce the prices of their products and services due to strong competition.
Nowadays, it is almost impossible to improve performance without incurring additional costs. The optimal way to maximize the economic result and effectiveness of a company is to reduce costs. Therefore, the company management needs to be able to manage costs effectively and properly to improve the competitiveness of the company. If company management has the ability of proper management, companies can achieve better economic results and tend to be more resilient to possible business risks.
However, company management should be aware of the fact that such reduction must not be to the detriment of the service and product quality so that it does not negatively affect the satisfaction of the current and future customers. The process of reducing costs takes time; it needs to be elaborated, well planned and systematic, and all possible risks need to be considered before it is started. The biggest cost item of the company under review is customer service -transport, which accounts for 29.13% of all corporate operating costs. The company does not operate its own transport; it uses the services provided by external carriers. The second biggest cost item is customer serviceassembly, which accounts for 27.34 %. The company uses the assembly services provided by Czech and Slovak providers or private carriers. The third biggest cost factor is rent (for renting warehouses, offices, equipment, vehicles, tools, etc.), which accounts for 15.52%.
The main aim of this paper is to perform an in-depth analysis of the current state of the company and subsequent efforts towards the reduction of the operating costs of transport while applying relevant rationalization solutions.

LITERATURE REVIEW
When dealing with the issue of cost items and their rationalization, the sub-objective for the purposes of this paper was to analyse relevant literary resources. Within the literary research, the authors focused on scientific papers that deal with the application of pair wise comparison and the methods of multi-criteria analyses in the context of transport and logistics. In [1], the authors analyse and apply the methods of multi-criteria analysis. They combine two variants, stochastic and deterministic, in the context of transport. In [2], the authors apply the TOPSIS method and provide two scenarios for solving a specific type of decision-making problem with multiple attributes while evaluating the results on the basis of several approaches to determine the usefulness of the approximation spaces λ. A modified TOPSIS approach implemented in decision-making problems is addressed in [3], where the authors broaden the current knowledge of the issue under review by applying a supply partner selection framework for ongoing public procurement. Specifically, the quantification of criteria weight is carried out using the fuzzy AHP method, while the TOPSIS method is used for the classification of the considered variants of supply partners.
Agarski et al. compare efficiency and safety optimization using multi-criteria decision analyses [4]. The research includes examination of the application of multi-criteria methods for determining the working procedures to streamline work efficiency and safety. For this purpose, a set of economic, technical and safety criteria was determined, and a number of scenarios were compared. Group decision-making of the selection of suppliers is described in [5], which discussed the development of a progressive approach to making decisions on sustainable determination of suppliers using the TOPSIS method in an interval-valued Pythagorean fuzzy environment. The article [6] presents several multicriteria decision analyses on specific examples. The authors describe the implementation of the AHP, WSA and TOPSIS methods in the location and allocation problems in practice. The evaluation of the supplier quality and transport system user equivalence is specified in [7], where the authors emphasize the definition of the basic approaches and instructions in accordance with the supplier quality evaluation. The manuscript [8] deals with transport interactions and the implementation of multicriteria decision-making.
Another study [9] proposes a credit scoring model using the hybrid analytic hierarchy process for determining the order of priority on the basis of the similarity to an ideal solution (AHP-TOPSIS). The implementation of integrated management systems, which has become a prerequisite for any organization that aims to become a competitive and sustainable company, is addressed in [10], which is directly related to the paper submitted. The objective of the paper [11] is to propose a model of selecting sub-contractors that would fully consider the influence of company requirements on the selection of sub-contractors. The aim is to understand the process of converting specific requirements of the company into evaluation criteria and calculation of weights. Providers of logistics services may play an important role in supply chain management when maintaining customer satisfaction and reducing the costs of supply chain management. This issue is addressed in [12], where the aim is to apply the approach of the fuzzy-analytic hierarchy process (AHP). Similarly, Phruksaphanrat and Borisutiyanee address the issue of company and supply chains using AHP [13].
Proposal of specific measures in the context of… 123

METHODOLOGY
As mentioned above, a specific provider out of three defined providers will be chosen using opted multi-criteria evaluation instruments, namely, the AHP and the Base-criterion methods, when determining the weights of criteria by the Saaty method of quantitative pair wise comparison and the Fuller pair wise comparison method.
As for the Saaty tool, decision-makers compare each pair of criteria and determine the values of preferences between each other. Subjective assessment of the investigators is partially eliminated by normalization of the geometric mean. Therefore, use of this method for this type of task is suitable. TheFuller method (also known as the Fuller triangle) is one of the techniques used for comparison of criteria as well. The criterion being more important than another one is explicitly marked with respect to all the pairs. Even though it lacks the exact value of significance given to criterion preference, using this method for this type of tasks is suitable.
AHP can effectively handle both qualitative and quantitative data to decompose the problem hierarchically, wherein the problem is thoroughly broken down. Using subjective ratings of pairwise comparison, this tool then assigns quantitative characteristics to each element, indicating their importance. This method does not require the exact values of individual criteria assigned to each variant to determine the overall variants' ranking. Its goal is to select the variant that results in the greatest value of the objective function. This is considered a compensatory optimization approach. In line with the aforementioned, application of the AHP method for this type of tasks is suitable. Last but not the least, the Base-criterion method is suitable for this type of task, given that it entails a general multi-criteria evaluation procedure of variants, which is, in typical cases, assigned for growing or declining priorities of variants considered. Its crucial objective is to compare individual values of variant consequences with values of the so-called Base criterion. In our case, the Base criterion is deemed as a target, in most cases with the most appropriate results (values) of all the criteria taken into consideration.

Saaty method of quantitative pair wise comparison
This instrument is based on pair wise-comparison of individual criteria when using Saaty 9-point comparison scale. The points presented in Table 1 represent the relationship between two criteria [14]. Table 1 Saaty 9-point comparison scale [14] Number bodů Descriptor 1 Criteria are equally important 3 Weak importance of the 1 st criterion 5 Strong importance of the 1 st criterion 7 Demonstrated importance of the 1 st criterion 9 Absolute importance of the 1 st criterion In addition to points 1, 3, 5, 7 and 9, the intermediate steps 2, 4, 6 and 8 can be used. The comparison of criteria pairs is described using the Saaty matrix = ( !" ) in Fig. 1.
If the j-th criterion is preferred over the i-th criterion, the inverse value is entered into the matrix. On the diagonal of a particular matrix, the value 1 is always entered because the criterion is equivalent to itself. It is an n · n matrix (a square matrix), and the elements of the matrix Sij are expressed by the estimate of the share of the i-th and the j-th criterion as expressed in formula (1): where i, j = 1, 2, …, n.

124
O. Stopka, V. Ľupták, I. Rybicka For the elements of the matrix S, it holds that, see formula (2): However, the elements in the matrix may not be consistent. The consistency of the matrix can be determined by the consistency index using formula (3) [15,16].
(3) where lmax is the largest eigenvalue of Saaty matrix and n is the number of criteria The consistency of the matrix S is sufficient if IS <0.1.
Weights vj can be calculated in several different ways. The most commonly used method is the weight calculation using the normalized geometric mean of rows in the Saaty matrix, which is referred to as the method of logarithmic least squares. The value of gi is calculated using formula (4) as a geometric mean of matrix rows [17].
where , = 1, 2, …, gi is the geometric mean, Sij represent values of the matrix and n is the number of criteria.

Fuller method of pairwise comparison
The Fuller method of pairwise comparison is used to determine the number of priorities for each criterion in relation to the other criteria in the set. The method is used if the ordinal information expresses the relationship between each pair of criteria being evaluated. Every two criteria are mutually compared and the number of comparisons (N) is thus determined as follows, see formula (6) The comparison is carried out using the so-called Fuller triangle, where the criteria are assigned order numbers 1, 2, …, n. The triangle is presented in Table 2 [18].
For each pair, a criterion that is considered the most important is selected. The number of selections of the j-th criterion is indicated as nj, and the weight of the j-th criterion is calculated according to formula (7) [18].
The disadvantages of this variant include the fact that the value nj for the least preferred criterion is 0. However, value 1 can be added to each resulting value nj, thus avoiding the zero-resulting value of the criterion [18].

AHP method
The method of the Analytical hierarchy process was proposed by professor Saaty. Its main purpose is to provide a framework for the preparation of efficient decisions in complex situations, and it also helps to simplify and accelerate the natural decision-making process. The method creates a certain hierarchical system of the problem, and as a decomposition method, it decomposes complex non-structured situations into more simple components. The Saaty method of quantitative pairwise comparison is used at all levels of the hierarchy structure, which are arranged in the order from general to specific elements. The more general the elements occurring in relation to the decision problem, the higher they are in the hierarchy. The higher level of the hierarchy includes one element only, which defines the objective of the evaluation. The element can thus be assigned the value 1, which is then divided among the elements at the second level of the hierarchy. Similarly, the value of the individual elements is divided at the other low levels until the evaluation of the lowest level, i.e., variant, is obtained [19]. Table 2 Scheme of the Fuller triangle The basic simple task of the Analytical hierarchy process includes3 levels: 1. evaluation objective, 2. evaluation criteria and 3. evaluated variants.
In more complex problems, the so-called sub-criteria level occurs between the criteria and variants. As can be seen in Figure 2, if two or more evaluators deal with the evaluation of the problem, another level of evaluators (experts) appears between the objective and the criteria. Their weights reflect the degree of its soundness [20].
The relationships between all components at the hierarchy levels are determined similarly as in the case of criteria. In a four-level hierarchy, i.e., when determining one objective, h experts, n criteria and m variants, the second hierarchy level contains one pairwise comparison matrix with the dimensions of h×h. For the third level, it is h matrices with the dimensions of n×n, and at the fourth level, it is n matrices with the dimensions of m×m.
In simple terms, using these calculations of matrices, the variants divide the weights of the criterion for which the matrix is constructed. If a sum of given values for all the criteria is calculated for each variant, its evaluation is performed in terms of all the criteria that create a basis of the resulting order of variants. The results of the AHP method are verified using the Base-criterion method (see the following section) [21].

Base-criterion method
The base-criterion method is suitable for the given problem mainly due to the fact that it is a technique of multi-criteria evaluation of variants, which is usually intended for growing or declining priorities. Furthermore, it is a method based on the maximization of utility. Its main principles include, in particular, comparison of the values of variant consequences with the values of the Base criterion. In this case, the Base criterion is perceived as a goal, in most cases, with the best results of all the criteria being evaluated [22].
The overall evaluation of the variant x, which is described by the vectors x1, x2, …, xm of the measured criteria values K1, K2, …, Km, is given by formula (8) when this method of multi-criteria evaluation of variants is applied. Weights vj can be determined, e.g., by the Saaty or the Fuller method [23].

126
O. Stopka, V. Ľupták, I. Rybicka Fig. 2. Hierarchy of the AHP method [20] ( ) = ∑ " · " C " D, ∑ " = 1, " ≥ 0, = 1,2, … , where the j-th partial evaluation of the benefit criterion (with growing priority) is given as follows, see formula (9): Moreover, for the partial evaluation of the cost criterion (with declining priority), it holds true that, see formula (10): where the vectors $ 3 , % 3 , … , 1 3 represent the Base-criterion selected as a vector of the best or predetermined values of the criteria in the set. The method represents the basis for performing certain partial evaluation of variants with respect to individual criteria by comparing the values of all variant consequences with the values of the Base criterion [24]. Table 3 presents the basic data important for deciding on the most suitable supplier for the company. The method of securing against defaulters or non-payers is intentionally not mentioned, since, in terms of the needs of the company, this information is not relevant. The price for the databank services is a price for inland transport and transport to the Slovak Republic (this is sufficient for the purposes of the paper). The remaining two providers offer uniform prices for transport across Europe (i.e. e.g. shipments to Germany, Hungary, etc. can be entered). All three databanks offer public databases for their users with android access. Table 4 shows the criteria based on the data in Table 3. They are a basis for making decisions on the selection of the most suitable provider. Criterion K5 is not numerical. It will be quantified as follows: if the databank does not have the possibility of mutual evaluation of clients using its services, the value of the criterion is 0. If the databank offers this service, the criterion is assigned the value 1.

Determining criteria weights for the application of the multi-criteria evaluation of variants
For the Saaty method of quantitative pair wise comparison of criteria, to enable the application of the AHP method and the Base-criterion method, weights of given criteria in the Saaty matrix were determined based on consultation with an expert team. As the first step, it is necessary to specify the relationship of each pair of criteria being evaluated, in which the degree of importance on a point scale of 1-9 is calculated. In terms of criteria weights, all input values can be obtained through a survey, i.e., a survey of expert opinions needs to be conducted. K4 number of users currently present in the database of the transport databank K5 possibility of mutual evaluation of transport databank clients (contractors and carriers) Source: Authors Table 5 below presents the information on experts who participated in the survey in terms of the definition of a set of criteria as well as the evaluation of their weights (input weights of compared pairs of criteria).
As for the Saaty matrix, the factors evaluated by experts were scored, and the arithmetic means of all sub matrix values were calculated. The individual values of the Saaty matrix were implemented in the process of gradual quantification, i.e., determining the geometrical mean and calculation of the resulting vector of criteria weights. The results presented in Table 6 show that the largest weight was calculated for K3, which was followed by K2, K1 andK5. The smallest weight was calculated for K4 [25].
Fuller method of pair wise comparison of criteria-the method of the Fuller triangle is based on priorities, where the list of criteria was presented to the above-mentioned expert team, who described the requirements for their prioritization and determined the input weights between the pairs of criteria being compared [26]. Table 7 shows the calculation of specific criteria weights. Table 8presentscriteria with their weights. The results, i.e., the weights obtained using the Fuller method, are, in terms of importance, the same as in the case of weights obtained using the Saaty method. Since criterion K4 has a zero weight, the weight of each criterion was increased by 1. The largest weight was determined in the case of criterion K3, followed by K2 (lower importance), K1, K5 and K4 (the smallest weight). The calculation using the AHP method and the Base-criterion method is based on the values presented in Table 9.

Results -Method of Analytic hierarchy process (AHP)
The first method to determine which of the three databank providers is the most suitable in terms of the management requirements is the method of the Analytical hierarchy process (AHP). The allocation of ratios using the Saaty scale was discussed with the above-mentioned expert team. Partial results of Saaty matrices are presented in the graph of AHP in Figure 3. By summing up the partial results on the left in the arrays P1, P2 and finally P3, it is possible to calculate their resulting value with weights obtained through the application of the Saaty method [27]. Summing up the values in the right column provides the final result for the calculation with weights obtained from the Fuller triangle.
The following Table 10 summarizes the results obtained on applying the AHP method.

Results -Base-criterion method
The application of this method is based on the matrix presented in Table 11. The first two criteria are minimizing (prices); the remaining three criteria are maximizing.
The resulting matrix is presented in Table 12. The column by SM shows the calculated resulting value for the criteria multiplied by the weights obtained using the Saaty method, while the column by FT shows the results calculated using the weights obtained using the Fuller method. It is obvious at first glance that the variants are non-dominated.
The order of variants (providers) in Table 13is the same as that in the case of the AHP method. The best option appears to be Trans.eu, whose result when calculated with the weights obtained using the Saaty method is 0.0793 higher than the result of RAALTRANS. Transpen rated third, with a difference of 0.0955 compared to RAALTRANS. The calculations using the weights obtained using the Fuller method confirm the results obtained through the application of the AHP method. Based on the results, Trans.eu is a compromise variant, followed by Transpen and RAALTRANS.

DISCUSSION
Based on the application of the methods, it can be concluded that the best option for the company is to use the services provided by the transport databank of Trans.eu, s.r.o. This compromise variant ranks first in the application of all methods. The ranking of the other two variants is not clearly defined, but as in the case of calculations that included weights obtained using the Saaty method, the difference is less significant than in the case of the Fuller method. We recommend to apply ranking based on the second type of weights (i.e., Fuller method).
If follows from the findings of the AHP method that the compromise variant significantly exceeds the values of the other two providers in the case when the calculations included the weights obtained using the Saaty method, with the difference between the best variant and the second best one being 0.0793. In the case of using weights obtained using the Fuller method, the difference is 0.4069. It can thus be concluded that the best variant is Trans.eu, s.r.o.
The application of the Base-criterion method confirmed the rating provided by the AHP method. Using the weights obtained using the Saaty method, the difference between the best variant and the second best one is 0.0793; in the case of the Fuller method, it is 0.4069. Thus, it is recommended to cooperate with the Trans.eu, s.r.o. company in the future.

CONCLUSION
A step towards rationalization of costs was the selection of one of the three providers of transport databank services using multi-criteria evaluation of variants, specifically the AHP method and the Base-criterion method. Using transport databank services enables the company to streamline its transport costs. The criteria weights were determined using the Saaty method of quantitative pair wise comparison and the Fuller method of pair wise comparison of criteria. Based on the method applied, it can be concluded that the most suitable provider of databank services for the company is Trans.eu, s.r.o., which was rated first in all the cases. The ranking of the other two variants is not clearly defined, however since the value-difference between the criteria weights obtained by the Saaty method 132 O. Stopka, V. Ľupták, I. Rybicka is not as clear as the weights of the criteria determined by the Fuller method, it is recommended so that the company under investigation takes into account the weights calculated by the Fuller triangle.
From the point of view of the given company, it is not possible to find out how many shipments are carried out via transport databanks entered into the portal by two carriers. However, they demonstrably use the services of the RAALTRANS databank. The proposal for the rationalization of costs consisted of selecting one of the three providers of databank services. Since at least two major carriers use these services, charging the company an average additional fee of 1,220 CZK per carriage, the company was recommended to use the databank services itself. The most suitable provider was selected using the method of multi-criteria evaluation of variants, specifically the AHP and the Base-criterion techniques. The criteria weights were calculated using the Saaty method of quantitative pair wise comparison and the Fuller pairwise comparison of criteria. As for the AHP method as well as the Base-criterion method, both types of weights were considered for verification. The best option turned out to be Trans.eu, whose use, calculated with the weights obtained using the Saaty method, is 0.0793 higher than the value of RAALTRANS. Transpen rated third, with a difference of 0.0955 compared to RAALTRANS. The calculation considering the second type of weights confirmed the results of the AHP method. Trans. EU was confirmed to be a compromise variant, followed by Transpen and with RAALTRANS rated third. In this specific case, 5 criteria were considered.
Based on the results obtained, it can be stated that the application of the AHP and the Basecriterion methods showedTrans.eu to be the most suitable option on the basis of a specified set of factors. It also can be declared that no similar publication that deals with analogous subject as this publication does has been presented yet. The submitted scientific paper adds value to the knowledge of proper determination and selection of transport companies in the Czech Republic and abroad while emphasizing the importance of defining corresponding criteria. The above findings present aspects in which the novelty and innovative solution of this research work lie.
Further research on the presented subject could focus on extending the compared set of criteria or extending the individual criteria by sub-criteria, or the application of other Operations Research methods, with a focus on the aspect of economic evaluation of the benefits achieved, which has not been addressed in our research.