Analytic network process in traditional market solid waste management in Malang Regency, Indonesia

One type of decision support process that assists in decision-making during Multi-Criteria Decision Making (MCDM) is the Analytic Network Process (ANP). ANP can determine the best pattern of sustainable traditional market solid waste management in Malang Regency. The six categories used in ANP analysis consist of environmental impacts (DL), technical operations (TO), regulations (PH), institutions and organizations (KO), financing (P), and community participation (PSM). The weighting results of the criteria that had the largest weight was the sub-criterion of Strengthening the Active Role of the Community (PSM1) with a limiting weight of 0.058271. The largest alternative weighting result was Scenario 3 (Integrated Solid Waste Management). Scenario 3 had a priority with the highest weight of 0.213951. The results of the sensitivity test when all criteria values were changed to 0.999 also changed the value of the alternatives. The changes in alternative weights when the sensitivity test was performed consisted of 3 sub-criteria that were sensitive to the changes, namely the sub-criteria of soil pollution (DL2), direct benefits (P4), and ease of operation (TO2). This research shows that the use of ANP provides a more efficient decision-making process in traditional market solid waste management.


Introduction
The Analytic Network Process (ANP) is a method for solving an unstructured problem that requires a dependency relationship between its elements.ANP is used to solve problems that involve dependencies between different alternatives or criteria [1,2].The ANP implementation uses multiple experts to analyze the relationship between criteria and the relative weight value between these criteria.By multiplying the criteria weights and alternative weights, a total weight for alternative priorities is obtained [3].This research aims to determine the best alternative for sustainable traditional market waste management in Malang Regency using the ANP method.
Commercial areas are the areas used as trade centers, markets, shops, hotels, offices, restaurants, and entertainment venues.Traditional markets include commercial areas that produce solid waste similar to household waste.The solid waste generation, composition, and characteristics are different from household waste so it needs its own management.The market in Malang Regency is an area where goods are bought and sold with more than one seller.It can be referred to as a shopping center, people's market, shop, mall, plaza, trade center or other designations.Traditional markets are managed by small and medium-sized traders, community selfhelp, or cooperatives and micro, small, medium enterprises (MSMEs) with the process of buying and selling goods through bargaining.Traditional markets in Malang Regency are divided into 4 clusters based on their size, services, and functions.

Research Method
In this research, several types of data are used, including data on criteria affecting the selected scenario, data on interdependent relationships between criteria, and pairwise comparison data between groups/criteria [4].The determination of the interrelationship between sub-criteria is obtained through a questionnaire of interrelationships between predetermined sub-criteria.The respondents consisted of stakeholders at the manager level/head of the traditional market management unit in Malang Regency.There are 33 markets in Malang Regency divided into 4 classes.The samples of this study were 3 markets representing each class, making a total of 12 samples [5].The questionnaire results were integrated from both groups of respondents and used to determine the block matrix linkage relationship.

Results and Discussion
Data processing was carried out according to the stages of the ANP method for selecting Alternatives using Super Decisions Series 3.2 software [6].The advantage of this ANP method is that the sub-criteria of one criterion can influence the sub-criteria of other criteria.ANP network models can be created before weight values are obtained.Criteria, sub-criteria, and Alternatives are constructed in a network to describe the relationship between criteria and Alternatives.The respondents consisted of stakeholders [7], and as many as twelve respondents (N) represented stakeholders from the solid waste management market.The recapitulation of the determination of the interdependency relationship between criteria was then calculated.If in a cell the number of respondents who chose was ≥6, it was concluded that there was an interdependent relationship between criteria.The results of data processing for each respondent were then compiled in Minitab software [8] to obtain an average value that represents the overall opinion of the market waste management.The criteria used in the study consist of environmental and nonenvironmental criteria [9,10,11,12], adapted to the conditions of the Malang district, which has many traditional markets, as shown in Table 1.Network ANP in Super Decision software is made according to the ANP network design, which is designed to show the dependency between sub-criteria and Alternatives.The network in Super Decision software is organized with several clusters.Clusters at the first level are the objectives of the ANP processing: selecting the best scenario for traditional market solid waste management.Clusters at the second level are the criteria for selecting the best scenario.There are six clusters, which means there are six criteria for selecting the best scenario, as each cluster represents one criterion, as shown in Figure 1.The marginal weighting value is obtained after obtaining the value in the matrix.The limit is a matrix that is squared until it reaches stability, where the value of each column is equal.Limiting is done by squaring the super-matrix until the columns have the same value and the sum of the weights of all nodes is 1.The limiting weight is also used to rank the sub-criteria so that it is known which sub-criteria have the greatest and least influence.The weighting results for all sub-criteria and alternatives and their rankings are shown in Table 2.The weighting result of the criteria with the largest weight is the subcriterion Strengthening the Active Role of the Community (PSM1) with a limiting weight of 0.058271.This shows that the active role of the community in traditional market solid waste management has a great influence on the decision making with ANP.The subcriteria air pollution (DL3) and community acceptance (PSM3) are ranked second and third with limiting weights of 0.057411 and 0.041267, respectively.
From the alternative weighting results, it can be seen that the highest weight of alternatives is in Scenario 3. Scenario 3 is considered better because it has a priority with the highest weight of 0.213951.The synthesis results of Super Decisions will obtain a priority order from the largest to the smallest value of the alternative scenarios, as seen in Table 3. Sensitivity tests are performed to analyze the stability of alternative priorities by simulating variations in the priority of criteria in the model.Special attention is paid to see whether the change changes the order of alternatives or not.For some criteria, tests were conducted by changing the weight of the value to find out the most critical criteria for changing the weight value on the four alternatives.From the prioritization results for the value of each criterion and alternative, the largest weight is currently on alternative Scenario 3 (Integrated Solid Waste Management).
Sensitivity tests were performed by changing each value on the criteria.An example of a sensitivity test on sub-criterion TO2 (ease of operation), the choice will change to Scenario 3 and 1 scenario 0 when the weight of the sub-criterion is increased to around 0.999.Thus, it can be said that the confidence interval of the decision to choose Scenario 3 is when the weight values of sub-criteria DL3, TO2, and P4 are in the interval 0<x<0.999.
The most sensitive sub-criteria are tested by attempting to change the weight value of the sub-criteria to 0.999.If the alternative choice changes, then the sub-criterion is a sensitive sub-criterion.The results of the sensitivity test on alternatives show that if all the values of the criteria are changed to 0.999, the value of the alternative also changes.
The results of the sensitive interval point test on the Soil Pollution sub-criterion (DL2) show the same weight value in Scenario 2 and Scenario 3, namely 0.316.An increase in the parameter value >0.845 causes the best alternative scenario for the Soil Pollution sub-criterion (DL2) to be Scenario 2 (composting).A parameter value between 0 and 0.845 indicates that Scenario 3 (Integrated Solid Waste Management) is the best Scenario for managing the Soil Pollution sub-criterion (DL2).When searching for sensitive interval points, it is found that the alternative parameter values of Scenarios 1 and 3 have the same weight.
The results of the sensitive interval point test for the Ease of Operation (TO2) sub-criterion are at parameter 0.749.If the parameter value is increased to a higher value, Scenario 1 is selected.This means that if the parties that will be directly involved in the waste processing consider the ease of operation to be the most important thing, then Alternative 3 is not the best decision.If the directly involved parties think that the ease of operation (TO2) is in accordance with what was filled in the questionnaire, then Alternative 3 is still the best.Thus, by keeping the values of the sub-criteria parameters in the range of 0 to 0.749, the decision to choose Scenario 3 as an alternative is the best step in market solid waste management.
The results of the sensitive interval point test on the Direct Benefits sub-criterion (P4) show the same weight value between Scenario 1 and Scenario 3, which is 0.335.This means that for a parameter value >0.674, Scenario 1 is the best for the Direct Benefits sub-criterion (P4), and by keeping the parameter value between 0 and 0.674, Scenario 3 remains the best alternative for the Direct Benefits sub-criterion (P4).

Conclusion
The Analytic Network Process (ANP) synthesis results show the highest priority for the integrated solid waste management alternatives with a normal weight of 0.213951.Changes in the alternative weights when sensitivity tests are performed can be seen that there are 3 sub-criteria that are sensitive to changes, namely the sub-criteria soil pollution (DL2), direct benefits (P4) and ease of operation (TO2).Other sub-criteria, by giving extreme weight to the sub-criteria, will still show that Alternative 3 is better.

Figure 1 .
Figure 1.Network in the Selection of Alternative Scenarios for Traditional Market Solid Waste Management.Clusters have several nodes, which are sub-criteria of each cluster or criterion.Clusters at the final level are alternatives in the form of scenarios that can be selected for traditional market solid waste management.Each cluster in the ANP network is connected by top-down-bottom-up arrow lines and looping, as in the criteria, where one criterion affects and depends on other criteria.An example of looping is in the Environmental Impact (DL) criteria cluster, which means that the sub-criteria water pollution (DL1) is influenced by soil pollution (DL2) and air pollution (DL3), and the sub-criteria soil pollution (DL2) is influenced by water pollution (DL1) and air pollution (DL3).Circular arrows indicate the interdependence of sub-criteria within a criterion (internal dependency), while straight arrows indicate the interdependence of sub-criteria between groups (external dependency).The marginal weighting value is obtained after obtaining the value in the matrix.The limit is a matrix that is squared until it reaches stability, where the value of each column is equal.Limiting is done by squaring the super-matrix until the columns have the same value and the sum of the weights of all nodes is 1.The limiting weight is also used to rank the sub-criteria so that it is known which sub-criteria have the greatest and least influence.The weighting results for all sub-criteria and alternatives and their rankings are shown in Table2.

Table 2 .
Weighting results with super decisions

Table 3 .
Synthesized results of super decision alternative ranking