Decision-making is a process present in the day-to-day of organizations and, when it comes to complex and competitive environments, the need for quick and accurate decisions becomes even greater.
Multicriteria decision methods were developed for a better structuring of problems involving decision making based on a series of factors, whether quantitative or qualitative, and even conflicting among themselves.
Examples of these methods are ELECTRE (Elimination et Choice Translating Reality), MACBETH (Measuring Attractiveness by a Categorical Based Evaluation Technique), SAW (Simple Additive Weighting), SMART (The Simple Multi Attribute Rating Technique), TOPSIS (Technique for Order Preference by Similarity to the Ideal Solution), ANP (Analytic Network Process), and the method discussed in this article, the AHP (Analytical Hierarchy Process).
The Analytical Hierarchy Process method was first introduced by Saaty in 1971 to solve scarce resource allocation in military planning. The main characteristics of this process are summarized in figure 1.
Since its introduction, AHP has become one of the most widely used multicriteria decision-making methods for solving problems in different areas such as vendor evaluation and selection, project prioritization for resource allocation, IT system selection, hiring and employee promotions. people and military decisions about whether or not to build defense systems (Saaty, 2008). In addition to reducing the subjectivity of decisions, the method allows, in some cases, a reduction in the duration of the process. Almeida and Murakami (2005) cite an example of a reduction from almost 300 days to 35 days in the decision to purchase software.
The methodology steps will be illustrated through three examples of AHP application: selection of a Logistics Service Provider (LSP), selection of the location of an industrial plant and selection of a port terminal. The first is a fictional case and the others are real cases.
CASE 1: SELECTION OF A LOGISTICS SERVICE PROVIDER (PSL)
According to the Panorama of Logistics and Railway Operators (ILOS, 2012), Brazilian industries have been betting on logistics as a way to differentiate themselves with customers, also increasing the demands on contracted logistics service providers (PSLs). PSL selection processes are increasingly rigorous, in which issues such as their ability to meet customer needs and their reputation in the market are analyzed.
In this way, the process of evaluating and selecting PSLs can be considered a multicriteria decision problem, with quantitative and qualitative factors to be considered in the choice, involving the application of the AHP method. A survey carried out by Viana and Alencar (2011) on supplier selection methods shows that, of 56 cases investigated, 50% used multicriteria decision support methods and, of these, almost half (43%) suggest the application of the AHP method.
The example described below is a fictitious case designed to illustrate the steps of the methodology in selecting a logistics service provider.
It is possible to describe the steps of the process in six steps:
- Define the problem to be solved;
- Define the alternatives;
- Define the relevant criteria for the decision problem;
- Evaluate the alternatives in relation to the criteria;
- Evaluate the relative importance of each criterion;
- Determine the overall evaluation of each alternative.
- Define the problem to be solved
In this case, the problem to be solved is the selection of a logistics service provider. In this stage, the type of knowledge to be sought is also defined, it is important to understand the requirements of the logistics operation to be outsourced and to map the main user areas involved in the contracting process.
- Define the alternatives
To simplify the example, four PSLs will be considered as alternatives: PSL1, PSL2, PSL3 and PSL4. In a real case, it is likely that time will be needed to research and build a list of PSLs that provide the desired service.
- Define the criteria relevant to the decision problem
The relevant criteria for the decision depend on the operational and managerial requirements of the operation to be outsourced.
A survey conducted by the ILOS in 2011 with around 300 logistics professionals evaluated the perception of these users in relation to PSLs. Some aspects analyzed in the survey were excellence in activities carried out such as transport of closed and fractioned cargo, integrated management of logistical operations, storage, project development and door-to-door distribution; the use of IT tools such as WMS (Warehouse Management System), TMS (Transportation Management System), satellite tracking; and time to market (ILOS, 2012)
Another study carried out a survey of 26 relevant criteria for the application of the AHP in the selection of logistics operators. Examples of these criteria are costs, service quality, company reputation, long-term relationships, information technology capacity, geographic coverage, risk management and services offered (ÇAKIR, 2009). For this simplified case, we will consider three criteria and six sub-criteria, illustrated in figure 2.
The method allows viewing the problem in a hierarchical structure with the number of levels deemed necessary. The structure contains the objective at the top and the alternatives at the bottom, where these are linked to all sub-criteria (Figure 2).
- Evaluate the alternatives in relation to the criteria
In this step of the process, the alternatives are compared pair by pair in relation to each sub-criterion. More generally, each level of the hierarchical structure is compared pairwise with respect to the level directly above it, that is, the alternatives are compared with respect to each sub-criteria, the sub-criteria are compared with respect to its criterion, and the criteria are compared with each other. relation to the objective (Saaty, 2008). Table 1 shows the comparison scale used in the method, which determines which criterion or alternative is better and how much better it is in relation to the pair.
The pairwise evaluation of the alternatives in this case generates six comparison matrices as shown in Table 2. Decision makers should define how much one PSL is better than the other in terms of transportation costs and the other five sub-criteria. These comparisons help to reduce the subjectivity of the process, especially when it is not possible to measure the criteria. However, the method allows the use of quantitative data, when available, to evaluate alternatives instead of pairwise comparison. In the case of a PSL selection process, information could be provided by the PSLs themselves through a questionnaire. Therefore, let's consider that we have the quotation of transport costs according to table 3.
With regard to qualitative criteria such as the software and systems sub-criterion, it is possible to identify, also through a questionnaire, which systems are interesting for the operation of each PSL, and to calculate a score.
- Evaluate the relative importance of each criterion
In this step, the decision makers involved in the process compare the criteria and sub-criteria in the same way as shown in Table 2. In the example, it would be necessary to create four comparison matrices, three comparing the sub-criteria with respect to their criterion and one comparing the criteria with respect to to the goal. The process can be conducted through workshops where the areas involved must agree on the relative importance of each criterion.
- Determine the overall assessment of each alternative
The overall assessment of the alternatives takes place through normalized relative scores and comparison consistency calculations. It is possible to mathematically model this method in spreadsheets, but in this case the software Expert Choice, a decision support tool based on the AHP method. The software calculates and normalizes the relative importance of criteria and subcriteria, the relative scores of alternatives, and checks the consistency of comparisons. The example in figure 3 shows the example of the relative score of the alternatives in terms of storage costs.
In this case, PSL1 had the best score in the global assessment. Figure 4 shows the overall evaluation of each alternative and their performance in relation to each criterion. The software allows performing sensitivity tests of the results regarding the importance of the criteria, sub-criteria and alternatives.
The case demonstrated in a summarized way the steps of the AHP for the selection of a Logistics Service Provider. The topic becomes relevant not only for contracting companies, but also for LSPs as selection processes become increasingly rigorous and structured, increasing the need for LSPs to adapt to customer requirements and improve measurement.
CASE 2: INSTALLATION LOCATION
One of the applications of the AHP method is to assist in the selection of installation locations. In this real example, we wanted to define the best city to locate a company in the steel industry.
The first step was taken by the company's top management, which predefined eleven potential locations for the installation.
Once the locational alternatives were defined, the next step was to identify which criteria are relevant to answering the problem's question. Through field research, consultations with specialists and interviews with the company's top management, four criteria and ten sub-criteria were identified (Figure 5).
The score of each alternative in relation to the sub-criteria was made in order to objectify the analysis, aggregating content from external sources, tacit knowledge of specialists and results of cost calculations in a single scoring scale.
As the last step of the AHP analysis, the definition of the relative importance between the sub-criteria was made in order to consider points of view from different hierarchical levels of the organization. Through workshops carried out separately between the company's operational, tactical and strategic teams, where each criterion and sub-criteria were compared two by two, the relative importance given to them by each of the groups was arrived at. In the academic literature, there are several ways of combining different evaluations, in this case these results were combined through a weighted average, with weight defined by top management, generating a single and robust result that incorporated different experiences and tacit knowledge of representatives from all over the world. the company (Figure 6).
The final result of the study indicated the ranking of preferred locations as well as the relative score of each one of them. In addition, sensitivity analyzes were carried out on the variables with the highest degree of uncertainty in the study, such as changes in demand, commercial policy, tax benefits and the mix of inputs. These analyzes indicated great robustness in the results, since the first three locations remained in the first three positions in all cases.
The application of AHP in this case was very successful. Through the methodology it was possible to obtain internal consensus in the company, adding to the model quantitative and qualitative factors and tacit knowledge of the people involved in the project.
CASE 3: PORT TERMINAL SELECTION
Another example of AHP application is for choosing a port terminal from which to operate. This application was presented by ILOS at the “XVI International Logistics Forum” (ILOS, 2010).
In this case, there was the possibility of carrying out the operation from three different ports, here called Port A, Port B and Port C. The selection criteria were: Logistic Service Provider, Port Terminal, Location and Access and Terminal Capacity . The tree of criteria and subcriteria can be seen in figure 7.
To score the alternatives in relation to each sub-criterion, a response scale was created for each one of them. Thus, comparing the ports pair by pair in relation to each of the sub-criteria, the relative position of each port in relation to meeting each criterion was obtained.
The next step was to define the weight of each of the criteria and sub-criteria in the choice. Again, the pairwise comparison was used, arriving at the result shown in figure 8.
Thus, it was possible to obtain the ranking of ports for the operation. In this example, the total cost of the operation was not considered as a criterion of the AHP model, being added later in a Cost x Benefit analysis where the “Benefit” is the result of the AHP analysis and the cost was normalized in order to obtain a proportional comparison between the ports as shown in figure 9.
In this case, the port that proved to be the most advantageous to carry out the operation was Port A. Due to higher costs, Port B was in second position despite having obtained a better performance in the ranking of the AHP model.
Thus, it is observed that the cost component can be treated in two different ways, according to the weight it should have in the decision. The first is to include cost as one of the analysis criteria, as shown in Example 1. The second way is to perform a cost-benefit analysis where the cost of the alternative is compared with the AHP ranking.
CONCLUSIONS
Among the advantages of using the AHP are the ease of use and understanding of the tool by all the agents involved, making everyone understand the problem in the same way and allowing the visualization of the interrelationship between the factors. However, the main gain in using the method is the ability to generate collaborative decisions, making each participant feel responsible for the final result.
Failed attempts to apply the AHP have already been observed in some companies that gave an extremely high weight to the cost criterion, making the analysis of the other criteria irrelevant. Thus, it is important that, before starting the process, interviews are carried out with the company's decision makers to identify the level of relevance of the qualitative criteria in decision making and thus determine whether the method is the most appropriate to address the problem in question. .
Among the most common difficulties in using the method, we can mention the amount of work required of decision makers to determine all the necessary comparison pairs, which requires increasingly scarce time in these professionals' agendas. Another common difficulty is ensuring that all criteria represented are independent or, at the very least, sufficiently different to ensure that an alternative is not doubly benefited, or harmed, by a single feature.
Thus, given the numerous variables involved in the day-to-day decision-making of companies, the AHP method stands as an important alternative tool to approach complex problems in a structured way, generating consensus and robustness in decision-making.
BIBLIOGRAPHIC REFERENCES
Institute of Logistics and Supply Chain (ILOS). Reducing subjectivity in choosing logistics service providers: application of the AHP method in choosing ports. XVI International Logistics Forum. Rio de Janeiro, 2010 .
Almeida, Martinho Isnard Ribeiro de; Murakami, Milton. Strategic Approach to Decision Making Based on the AHP Method (Analytic Hierarchy Process). In: XVIII Latin American Strategy Congress, 2005, Santa Cruz de la Sierra. SLADE, 2005.
Saty, Thomas L. Decision making with the analytic hierarchy process. Int. J. Services Sciences, Vol. 1, No. 1, 2008.
Institute of Logistics and Supply Chain (ILOS). Panorama Logistic Operators and Railways: The Best in the Perception of Their Users, 2012.
Viana, Joana Coelho; Alencar, Luciana Hazin. Methodologies for selecting suppliers: a literature review. Federal University of Pernambuco, 2011.
Çakir, erdal. Logistics outsourcing and selection of third party logistics service provider (3PL) via fuzzy AHP. Master Thesis. Bahçesehir University, Institute of Science Industrial Engineering, 2009.
