Comparative evaluation of alternatives for management of wood wastes by using multi-criteria decision tools

In Turkey, the wood industry is a prevailing sector generating economically valuable by-products such as sawdust, wood chips and wood shavings. The utilization of these materials supports an important contribution to the revival of the economy. Selection of the best management alternative considering economic and environmental factors synchronously requires the application of statistical decision-making methods. The significance of this study is to determine the suitability of promising alternatives for the utilization of wood wastes (production of pellet, panel, cemented wood composites and pyrolysis) by using multi-criteria decision-making methods (PROMETHEE and FPROMETHEE). Super Decision Program was used to evaluate the importance of the selected criteria (environmental impact, operation risk, installation cost, applicability, operation cost and market competition) based on data obtained from the 9-scale questionnaire. In PROMETHEE and FPROMETHEE applications, both equally weighted and calculated importance of criteria was considered. According to results of AHP, installation cost was the most important criterion with 31% ratio. Results of the study showed that pellet production was the best alternative in all applications with the highest-ranking values. Wood-based panel production was followed by pellet production; however, it was seen that the production of cemented wood composites and pyrolysis are not suitable for the utilization of wood wastes.


Introduction
The limited world resources and the rapid increase in the production of waste have led to demand for reducing the amount of waste produced, therefore, reusing and recycling materials called the implementation of a zerowaste approach becomes a crucial topic.The zero-waste approach rather than waste disposal protects virgin materials on Earth and facilitates increasing the lifetime of the waste in the economic cycle [1,2].
Wood is one of the important raw materials that should be considered in the zero-waste approach [3].It is used as a solid fuel to supply energy as well as it is a fundamental raw material in sectors such as construction, furniture, and printing paper industry [4].Excessive consumption of woody materials causes the loss of forest areas and increases greenhouse gas emissions in the urbanizing world [5,6].According to FAO reports, a total of 178 million hectares of forest area on earth has been lost since 1990 and it is detected that 11.5 million hectares (50.32%) of the forest area in Turkey are reserved for economic functions and wood production [7,8].In Europe, about 33.7 thousand m 3 wood products became wood waste, 46% of which was recycled and 51% incinerated [9].Similar to the most of the developing countries, the high demand for wooden products for various uses leads to an increase in wood waste in Turkey.Wood-based panel (particleboard) production is the most prominent recycling option for wood waste.The amount of recycled wood waste used in particleboard production varies between 50% and 60% in developed countries such as the United Kingdom, Belgium and Denmark [9].whereas this ratio is approximately 10% in Turkey [10].Depending on the wood quality standards, there are also recycling options such as production wood chips, pellets and refused-derived fuel.Tian et al (2017) revealed that wood output is an important variable affecting wood consumption, and it has been reported that encouraging recycling in places where wood resources are abundant may be more effective in reducing wood consumption.Therefore, it has become a challenging issue to manage both resources and wastes of woody materials efficiently [11].Consequently, in order to increase recycling rates in developing countries, it is important to determine the best recycling alternatives for the region.
Decision making involves choosing some course of action among various alternatives.Qualitative or quantitative problem parameters must be evaluated together in decision-making processes.This makes MCDM a practical and comprehensive evaluation strategy in decision-making processes [12].In this context, multicriteria decision-making methods (MCDM) are beneficial decision-support tools for the integration of various technical information and a combination of stakeholder judgments [13].MCDM can be classified as intelligent and traditional methods.Neural networks as one of these intelligent methods attempt to simulate the human brain by collecting and processing data for the purpose of 'remembering' or 'learning' [14,15].These models are deeply concerned with the calculations of weights among nodes or require cumbersome computations to evaluate alternatives [16].Traditional methods such as analytic network process (ANP), Elimination Et Choix Traduisant la REalité (ELECTRE), VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR), decision making trial and evaluation laboratory (DEMATEL), the technique for order performance by similarity to ideal solution (TOPSIS), Preference Ranking Organization Method for Enrichment Evaluations (PROMETHEE), AHP (Analytic Hierarchy Process) are widely used in different problem areas [12].
The most important deficiencies of ANP method are ignoring different effects between clusters and the possibility of significant problems due to measurement error [17].ELECTRE produces rankings of each alternative but lacks objective data to provide knowledge about the differences between alternatives.In addition, the complexity of VIKOR method shadows practical implementation of the method [18].The Dematel method, on the other hand, can be difficult to collect accurate and reliable data on the relationships between factors, especially when the data is limited, incomplete, or uncertain.Insufficient or incorrect data can weaken the validity and reliability of an analysis [19].TOPSIS is another MCDM method based on Euclidean distance calculations which do not take into consideration the correlation of data.Therfore, the results are affected due to the information overlap [20].
Among various MCDMs, PROMETHEE, developed by Jean-Pierre Brans in 1982, is prominent due to features of simplicity, clarity and reliability of the results [21].Usage options such as preference-indifference make the use of PROMETHEE shine [22,23].It provides accurate determination of the sequencing of alternatives with linguistic expressions considering expertise, competence and experience of the decisionmakers.However, using linguistic expressions may cause ambiguity and blurring.In such cases, Fuzzy-PROMETHEE (FPROMETHEE) method using fuzzy numbers may be benefical [24].
In MCDM applications, the analytic hierarchy process (AHP) has been commonly used to determine the subjective weights of selected criteria [25].In AHP method, the consistency of the results can be measured by pairwise comparison.AHP allows the evaluation of multiple criteria, sub-criteria and lower-level criteria as well as their impact on the decision-making process [26].
As seen from recent literature, PROMETHEE and FPROMETHEE methods can be efficiently used for the evaluation of possible alternatives in waste management practices [27].Wu et al (2018) used TODIM method combined with PROMETHEE-II for ranking the alternatives to determine the plant site of waste to energy [28].Makan and Fadili, (2020) used PROMETHEE to evaluate different composting reactors considering environmental, financial/economic, social and technical criteria [29].In another study, the alternatives of food waste management (wet process, indoor composting, outdoor composting, anaerobic digestion, and transportation to the mainland) were evaluated considering criteria including environmental, social, and economic standards [30].In the study, the most appropriate waste disposal site was selected by a combination of PROMETHEE and AHP methods [31].The same combination was also used to determine performance indicators of all key components of municipal solid waste management systems [32].Seikh and Mandal (2023) expressed the linguistic expressions for evaluation of biomedical wastes with fuzzy numbers and applied PROMETHEE.In the study, the weighting of the criteria was carried out by applying the Stepwise Weight Assessment Ratio Analysis (SWARA).Bio-chemistry lab was chosen the best biomedical waste treatment in all scenarios [33].Liang et al (2020), investigated the selection of best hazardous waste disposal site using combined hesitant fuzzy PROMETHEE and AHP.It is emphasized that determining the weights by AHP method increases the accuracy and objectivity of the study [34].Hendik et al (2022), also performed Ranking PROMETHEE, PROMETHEE Rainbow, PROMETHEE GAIA, and PROMETHEE V for determination of red mud cluster sites .This study thought to be groundwork for future research to find a consensus among stakeholders interested in the management of waste streams from alumina pilot plants [35].Angelo et al (2017) used the software VIP-Analysis to determine the most appropriate decision for food waste management in terms of the Life Cycle Assessment.PROMETHEE methods not used only for solid waste problems in environmental researchs but also used varios parts of the environmental engineering tasks [36].Gichamo et al (2020) used FPROMETHEE to discuss natural waste water treatment alternatives.In the study, TOPSIS method was applied to validate the results of the survey [37].Özdemir et al (2020), compared efficiencies of ANP and PROMETHEE methods to evaluate the alternatives for leachate treatment.Although the order of alternatives has changed, both of the MCMD techniques indicated that the best treatment method was combined treatment with municipal wastewaters [38].
The major objective of the present study was to evaluate the appropriateness of prevailing alternatives (production of pellet , panel, cemented wood composites and pyrolysis) for the utilization of wood wastes.With this aim PROMETHEE and its combinations with AHP and fuzzy approaches were applied.Evaluations were made considering criteria such as environmental impact (EI), operational risk (OR), installation cost (IC), and applicability in Turkey (AT), operational cost (OC) and market competition (MC) of the composed product.The featured outcome of the study is efficient and fast decision making by taking account of linguistic terms.It is thougt that the results of the work will enlighten both the academic and sector audiences about selection of the best alternative in management of wood wastes.

Method
Defining the alternatives and criteria Wood wastes are known as having great prospects for efficient energy recovery, recycling and reuse [39].Production of wood-based panels (WBP), pellet (PP) and cemented wood composites (CWC) are preferred alternatives for the utilization of wood wastes [40,41].WBP are produced at dissimilar decomposition stages under compression by pressure with the addition of synthetic resins [42].PP involves drying, grinding and compressing stages, respectively.This product has a large international market that is comparable to biodiesel or bioethanol in terms of processing volume [43].CWC are smooth surface materials formed by combining wood chips, cement, water and chemicals in appropriate proportions [44].Typical structural properties of cement (resistant to water, moisture, burning and decay) and wood (lightness, elasticity and workability) provide unique material characteristics to CWC [45].Apart from the industrial utilization of wood wastes, pyrolysis (Py) is a promising recovery method based on thermal decomposition for converting wood biomass into alternative fuels [46,47].Energy and bio-oil are the main outcomes of the Py process [48,49].
In models based on MCDM, the aim is to promote an alternative between different priorities when calculating a solution.The optimal solution is the one performs well from all other alternatives when compared to the same set of priorities [50].The PROMETHEE method, on the other hand, evaluates all alternatives by pairwise comparison.Then, as a result of these comparisons, a score is formed based on pairwise comparisons [40,44,51,52].AHP was used to attribute weights for each criterion in order to reflect the effectiveness of the selection of the best alternative in the PROMETHEE method.In this study, alternatives, and decision criteria were selected based on literature research, the situation of the region and the opinions of experts.Criteria selection involves environmental impact assessment, evaluation of operational risks, proposing installation and operational costs, feasibility and market competition.An analytical description of each group of criteria was expressed on a scale from 1 (least favorable) to 5 (most favorable).
Application of MCDM methods consist of many different uncertainties in terms of method determination and selection of criteria weights [53].In the study PROMETHEE method was prefffered considering its superiority in evaluating various qualitative and quantitative criteria.Additionaly, above-mentioned literature put forwads remarkable efficiency of PROMETHEE in waste management applications.

Applied mathematical methodology
PROMETHEE method provides a selection of the best alternative among the possible ones based on the criteria proposed by the decision-makers.Alternatives are compared in pairs based on criteria with the selected preference function.The basic steps of PROMETHEE are given below [54,55].
Stage 1: Decision makers, alternatives and criteria are determined and criteria weights (AHP) are calculated.Stage 2: The most appropriate preference function should be selected in order to show the structure and interrelationship of predetermined evaluation criteria.There are 6 different preference functions (the usual type, U type, V type, level type, linear type and Gaussian type) used in PROMETHEE applications (table 1).
In this study, a linear preference function (5th type) was chosen for all criteria.This enables the decisionmaker to determine his preference for an alternative that is above the average.p was determined as '1' as the smallest difference value determined by the decision maker to create a definite preference.The indifference value of q was determined as '0'.
Stage 3: Based on the selected preference function, pairwise comparisons of the alternatives are made for each criterion according to the following expression: Where, a and b, are two alternatives.Correspondingly, preference indices are determined for each pair of alternatives: Stage 5. A partial order of alternatives is obtained with PROMETHEE I by comparing the positive and negative superiority values of the alternatives.This stage provides detailed knowledge about preference (superiority) of alternatives over each other, alternatives that are indistinguishable from each other, and alternatives that could not be compared with each other.Stage 6.Finally, a complete ranking of alternatives is obtained with PROMETHEE II.The exact priorities of the decision options are determined by the formula in equation (4):

FPROMETHEE methodology
In this study, the PROMETHEE approach with 5-point scale criteria was used to determine the most suitable alternative in the evaluation of waste wood sawdust.Since the decision-makers evaluated the criteria with subjective and linguistic expressions, it was thought that more sensitive results would be obtained by using fuzzy numbers instead of a 5-point evaluation scale, and FPROMETHEE analysis was performed.Rayanne et al (2021) also suggested using a fuzzy approach to incorporate uncertainty and reveal criterion weights.In the study, triangular fuzzy number equivalents of five linguistic variables proposed by Li (1999) were used [56,57].Table 2 Triangular fuzzy number equivalents of numbers expressed with five linguistic variables.
x q q x q p x q p 0, , 2nd Type (U Type) x q x q p q x q p x q p 0, , 1, q. Indifference Value.p. Definite Preference Threshold.s.Intermediate Value Between p and q.PROMETHEE and FPROMETHEE analyses were performed using Visual PROMETHEE software.For the PROMETHEE solution, the linguistic expressions of the decision-makers were scored according to the 5-point scale, their averages were taken for each criterion, and data entry was made in this way (table 2).For the FPROMETHEE analysis, the linguistic expressions of the decision-makers were expressed with fuzzy numbers and the averages were taken based on the criteria and clarified.The averages of the fuzzy numbers created with 5-point scoring for the PROMETHEE analysis are.'D' represented the clarified values: AHP methodology and weighting of criteria In the AHP method, the survey technique was used.The importance of each criterion was calculated and the Super Decision Software was used during the analysis.In the present study, weighted and equally weighted results were presented.Environmental impact, operating risk, installation cost and operating cost were set in the direction of minimization, while the applicability in Turkey and the market competition were set in the direction of maximization.
In Super Decision Software, a hierarchy model was created as the first step to determine the weights of the alternatives.Each criterion was compared in pairs considering the importance of the criteria.Weight factors were determined for each preference calculated using the pairwise comparison matrix.The matrix was filled using the AHP concept having scales with values from 1 to 9 to rate relative preferences for items.Based on judgmental preference, questionnaires were distributed to experts and all participant consent was gained for this study.Consistency and sensitivity analysis were performed to check the reliability of the obtained data [58,59].

Result and discussion
The results of the survey The survey investigated the importance of each criterion in terms of determining alternatives.The average results of the survey (PROMETHEE avg and FPROMETHEE avg ) and also DF Values avg were depicted in table 3. EI, OR, OC and IC were evaluated by minimization direction.WBP was the most applicable recycling approach for woody wastes in terms of EI with values of 1.83.Results for pellet production, CWC and Py were 2.00, 2.33 and 2.5, respectively.Pellet production was found to be the best alternative in terms of OR (1.67), IC (2.17), OC (2.33).Py was the most expensive alternative considering the highest values of IC and OC (4.17 and 3. 67, respectively).AT and MC were evaluated by maximization direct.According to AT values, the most favorable alternative was determined as pellet production (4.67).It was followed by WBP production with a value of 4.33.Py and cement composites were also interpreted equally with a value of 3.33.It was seen that pellet production and Py took the lead during the evaluation of the alternatives in terms of MC with a result of 4.17.
Wood waste can be considered for energy recovery.However, environmental risks may arise during the thermal process because of chemical contents such as oils, adhesives, paints and varnishes used in the fabrication of the furniture.Wood treated with resins such as urea formaldehyde, melamine formaldehyde and phenol formaldehyde shows higher decomposition temperatures [60].This situation may cause problems in the product quality and operating costs in pyrolysis.Survey results reveal that experts are concerned about woodwaste based pyrolysis.

The results of AHP
The weighting of the criteria was thought to be the crucial step for determining the best alternative because not all alternatives are of equal importance.Therefore, the weightings were calculated by Super Decision Software.Figure 1 showed that IC was the most important criterion in the determination of the best alternative for woody waste fate with a result of % 31.The weights of OC, OR, MC, AT and EI resulted in % 20, % 19, % 17, % 8 and % 5, respectively.
Promethee I partial sequence and Promethee II net sequence While performing the partial sequence of PROMETHEE I, positive and negative currents between +1 and −1 were calculated for each alternative.Positive currents indicated the superiority of one alternative over the others.In PROMETHEE II, a clearer ranking was obtained with the net currents resulting from the difference between positive and negative currents.Alternatives between 0 and +1 in PROMETHEE II were the alternatives that will be at the top of the ranking.Rankings for equally and calculated weighted criteria were presented in figures 2 and 3 for all applications, respectively.In this study, all tested MCDM resulted similar order of alternatives.Determination of the weights of the criteria provided more accurate results in PROMETHEE analysis.PROMETHEE II net rankings of PP, WBP, CWC and Py were 0.67, 0.47, −0.51, −0.63, respectively in PROMETHEE whereas in weighted PROMETHEE these value s were calculated as0.73, 0.41, −0.46 and −0.68, respectively.
The differences between the scores of PP and WBP evaluated on the weighted criteria were higher in this study and PROMETHEE combined AHP method showed that PP is the alternative to be selected with a higher score.Furthermore, this method's results (−0.46) showed that CWC is decided to be a little bit better alternative if it is compared to the PROMETHEE method (−0.50).On the other hand, it has been deducted that Py could be a much worse alternative.In FPROMETHEE and combined AHP and FPROMETHEE, the results were very close to each other.In this study, there was no difference in the choices expressed with fuzzy numbers and calculated in this way.It is concluded that the reason for the difference was only in the expression of fuzzy numbers with 3 and taking values less than 1.Considering that the PROMETHEE method could be calculated both by expressing it linguistically and by presenting numerical data by analyzing the results, it was predicted that using data expressed with linguistic expressions in FPROMETHEE gave more accurate results in the analysis.Related literature has reported that PROMETHEE method can also provide successful results in nonnumeric expressions [61].PP was determined as the best alternative in all analyses.WBP production and PP alternatives have taken close values in the analysis.Hossain and Poon (2018) comparatively evaluated the life cycle analysis of wood waste management strategies generated from construction activities.Three alternative scenarios, including the recycling and reusing of wood waste to produce organic polymer based particleboard, cement-bonded particleboard and energy generation from bio-fuel (wood pellets) were compared with landfill disposal.The results showed that energy generation from wood-pellet was the best strategy.Results of the present study is also compatible with the life cycle analysis results in the literature [62].
The demand for biomass used for bioenergy purposes has been increased due to the concerns on climate change and energy supply security.Additionaly, as mentioned in policies targeting renewable energy generation, large amounts of solid biomass is required for industrial uses, especially in the European Union (EU-27).Therefore, there is a rapid increase in trade of of solid biofuels, mainly wooden pellets [63].
In this study, the most appropriate alternative was determined to be PP for wood waste management..It was declared that the use of wood waste in the panel industry will both produce high-quality boards and provide ecological and economic benefits [64].Py and CWC had lower positive current values compared to other alternatives.Although some recent studies indicated that the composition of food and wood waste is promising alternative for energy utilization [65] this bio-fuel may cause several operating problems (heating rate, catalystto-biomass ratio, the reaction temperature, the residence time of the biomass feedstock, the vapor residence and, alkaline ash production) during incineration [66].

The GAIA plane
The Geometrical Analysis for Interactive Aid (GAIA) plane referred to a visual and interactive procedural representation as an extension of PROMETHEE results [54].On the GAIA plane, alternatives and criteria are shown as points and vectors, respectively.The decision line shown with the red vector indicates the most suitable alternatives for the decision-maker.Belgin and Balkan (2020) used GAIA plane as performance evaluation technique in PROMETHEE analyses [67].In the present study, this method was performed for PROMETHEE and FPROMETHEE with AHP and the results were given in figure 4. As seen from the figure, PP was the best alternative, which was also confirmed by the net ranking results of PROMETHEE II.Since the vectors showing market competition, applicability and operational risk criteria in Turkey were long, this indicated that these criteria have great importance in influencing the decision bar.Vectors pointing in the same direction belonged to criteria with similar properties and criteria bars showing different directions belonged to conflicting criteria [68].The PP alternative, which stands out among the alternatives in the full ranking in the GAIA visual analysis, was positively affected by operating cost, operating risk, installation cost and applicability criteria in Turkey.

PROMETHEE rainbow
PROMETHEE rainbow is a PROMETHEE representation method used to highlight the positive and negative impacts of the criteria for each alternative that are shown as bar graph [69].Positive slices appeared upwards and correspond to positive traits whereas negative slices appeared downwards and represented negative points.This graph enables visualization of the characteristic profiles of all cases, taking into account the criterion weights [70,71].
In the study, the rainbow diagram was created to observe the impacts of criteria on the selection of the best alternatives (figure 5).As seen from the rainbow diagram, the criteria for PP and board production were all positive while criteria for cement wood composites denoted to be negative.For the Py alternative, only criterion of the market competition resulted positive.

Conclusion
In the study, PROMETHEE, AHP-PROMETHEE, FPROMETHEE and AHP-FPROMETHEE techniques were applied to evaluate different alternatives (PP, WBP, CWC and Py) for the utilization of wood wastes.
In PROMETHEE and FPROMETHEE methods, a ranking was made by accepting the criteria having equal importance.Furthermore, these methods were combined by AHP to determine the importance levels of the tested criteria.Numerical values for the importance levels were determined as 0.31, 0.20, 0.19, 0.17, 0.08 and 0.05 respectively for, installation cost, operating cost, operational risk, market competition, applicability in Turkey and environmental impact.Obtained results showed that installation cost has a relatively higher impact on determination of preference order.
Results of both techniques demonstrated that PP was the best choice among the tested alternatives whereas the panel production ranked second in all scenarios.Policies encouraging use of waste wood, which is below 10% for panel production sector in Turkey, may may yield displacement in the options.CWC was found to be less preferable indicating importance of new attempts providing improvement in its market share.
Results of the work indicated FPROMETHEE and PROMETHEE methods addressed the same preferences despite the different scores.The difference in preference scores was due to linguistic expressions being less than 1.Additionaly, obtained results demonstrated that combined AHP methods and visual PROMETHEE can be used to determine the best alternative in case of non-numerical criteria.If the data set was not very large and dispersed, PROMETHEE can also provide good results.Thus, it is possible to minimize the loss of time and labor.The novelty of this study is that it provides information on how to utilize sawdust waste, reflecting expert opinion in Turkey.The results obtained from the study will guide decision-making processes in wood waste management especially for developing countries.
This study was focused on the suitability of alternatives for the evaluation of wood waste within the framework of the possibilities in Turkey.Hence similar studies can be conducted for other countries to generalize the findings.The prevailing limitation of the study was absence of real-world data.The sole data source of the study was the surveys and this may cause a certain level of subjectivity.Future studies should be performed by using actual data sets.Furthermore, it is suggested to evaluate the similarities and differences between preferences of experts and life-cycle analysis of options.

K 1
In equation(2), W i represents the importance weight of the criteria and the preference index of a and b alternatives evaluated by k criteria with importance weights (I = 1, 2, ..,k).Stage 4: positive (Φ+) and negative (Φ-) superiority values are calculated for each alternative by equation (3):

Figure 1 .
Figure 1.The weights of criteria (%) determined by Super Decision.

Table 1 .
The preference functions of promethee approach.

Table 2 .
Triangular fuzzy number equivalents of numbers expressed with five linguistic variables.

Table 3 .
The average values of each criterion for both methods.