Introduction
In a scenario where the planet is still experiencing the consequences of a COVID-19 pandemic, in parallel with high fuel prices as a result of the war in Europe, the challenges of a competitive logistics operation demand more and more attention from companies.
In this context, where transport costs are on an upward trend and the consumer's purchasing power, harmed by inflation, is reduced, it becomes evident that the search for optimization in the logistics operation is critical. The growth of e-commerce has put pressure on the operations of many companies, and even those that managed to cope well internally with this pressure, by rescheduling production lines and recalibrating inventory levels, may be suffering from the last mile and with the need to serve a high number of delivery points. As an alternative to make it possible to meet this demand, many companies outsource the operation through logistics partners.
However, a complex step in the logistics outsourcing process is choosing which logistics operator to hire. According to research carried out by ILOS in partnership with ABOL (Brazilian Association of Logistic Operators), there are about a thousand logistics operators in the market. There are still 280 transport companies operating in the country, which allows for countless choices of partners. This fact reflects the need for a methodology that not only qualitatively assesses compliance with certain parameters such as cost-benefit, information flow and adequate structure, but that hierarchizes and seeks to quantify how well these companies meet these criteria. However, what brings complexity to the issue is that some of these parameters are in fact qualitative and difficult to measure. So how is it possible to quantify and rank qualitative data?
Weaknesses of the traditional AHP methodology
A widespread methodology to solve this problem is the AHP, in which values are established for each criterion comparatively, defines how much each operator meets each of them and then equates the problem generating a hierarchical definition of the analyzed options as a response. However, in the context of mathematics, this methodology is subject to some criticism, including:
- The AHP scale does not allow expressing a degree of uncertainty with respect to comparisons, and incomparable alternatives are not allowed. In addition, the need for consensus to determine weights and priorities can also be seen as a disadvantage, as it can lead influential leaders to distort the opinion of the rest of the team (KANGAS and KANGAS, 2005; VILAS, 2008).
- One of the biggest criticisms of AHP concerns the reverse rank, which can cause problems due to the fact that the insertion of a new alternative in a decision problem can radically change the previously established ranking (KANGAS and KANGAS, 2005; BOUCHER et al., 1997).
- The weights obtained from the pairwise comparison are strongly criticized for not reflecting people's real preferences (LINKOV and STEEVENS, 2008).
Therefore, such a method, although it is the subject of many discussions in scientific articles, proves to be not so efficient in dealing with the growing complexity of reality. Therefore, a methodology that incorporates uncertainty in its approach is necessary.
fuzzy logic
A fuzzy logic (also called fuzzy logic) proves to be a viable alternative to the weaknesses of the above solution, since it was developed precisely with the aim of incorporating uncertainties into the problem and doing their mathematical treatment. This happens through the treatment of set theory with the relativization between values through the degree of pertinence of an item to a universe. That is, a logistics service provider, in a traditional logic, would have the “Price” criterion evaluated as low, medium or high. Fuzzy logic understands that this “Boolean” approach is simplistic and considers that the same price can simultaneously belong to the “low” and “medium” sets. The image below seeks to exemplify this degree of belonging to different groups.
Figure 1: Example application of fuzzy logic. Source: ILOS
With regard to the methodology, it basically consists of 6 steps. First, the problem is defined, which, in this case, would be the selection of the logistics service provider within the strategy outlined by the company. Then, steps 2 and 3 are carried out (common to the AHP methodology): the elaboration of the criteria comparison matrix and the verification regarding the consistency of the comparisons. Afterwards, the fuzzification of the values is carried out by establishing vector values for each comparison and calculating the weights of each fuzzy vector given by the alternatives of logistic operators. Finally, the values are defuzzified and a ranking is established from the highest to the lowest of the operators that best meet the defined criteria.
Figure 2: Fuzzy logic application steps. Source: ILOS
Conclusion
Therefore, the methodology is evidenced as an alternative to contemplate the complexities of the real problem. Although it is not widely disseminated as a common practice in the market, the graph below reveals that in academia, for cases of supplier selection, the “newborn” fuzzy approach was the method most chosen as the subject of scientific articles between 2002 and 2011 by configure itself as a solution.
Figure 3: Ranking of MCDM methods chosen as the subject of scientific articles between 2022 and 2011 . Adapted: Lima Junior, Osiro, Carpineti, 2013.
It is worth mentioning some weaknesses of the method, which is still being established in the market, but which continues to be the subject of several articles and dissertations. First, it depends on an adequate choice of which criteria will be used to evaluate the alternatives, in order to avoid overlapping concepts. In addition, the choice of evaluators is also a determining factor for the success of the methodology, especially in cases where few people will be involved in the evaluation, since the score of each alternative may be biased by a small sample.
O 28th International Supply Chain Forum, which will take place between October 18 and 20, 2022, in person and online, will feature Brazilian and international cases to discuss the best practices and solutions currently adopted by major players in the logistics market. For supply chain decision makers, it's an excellent opportunity and worth participating!
References:
– ABOL (02/06/2022) – 69% of logistics operators were unable to pass on high costs in 2021, says Ilos/Abol
– Multi-Criteria Decision Analysis: Environmental Applications and Case Studies – Nanomaterials: Background and Environmental Challenges
– Néstor Fabián Ayala and Alejandro Germán Frank (June/2013) – Multicriteria analysis methods: a review of strengths and weaknesses
– Fabio Luiz Peres Krykhtine, Antonio Carlos Dias Morim, Natália Gonçalves Pires do Vale, Luiz Eduardo Netto Sá Fortes and Armando Gonçalves Celestino Neto (Oct/2013) – Applying Fuzzy Logic in a Multicriteria Selection Model for Multiclients
– Francisco Rodrigues Lima Junior, Lauro Osiro and Luiz Cesar Ribeiro Carpinetti (2013) – Multicriteria decision methods for supplier selection: an overview of the state of the art
