Application of MCAT to provide multi-objective optimization model for municipal waste management system

Abstract

Choosing an appropriate municipal waste management method is a very complicated environmental problem in cities. This research introduces an optimization model for waste management in the southwest region of Tehran province. It was developed by a metaheuristic algorithm that was used to minimize the economic and environmental costs. Incineration, composting, recycling and landfilling waste management methods were considered. Three scenarios were developed to determine the optimum allocation of waste to each method such to fulfill the objective of overall minimum of environmental burdens and costs. A multi-objective scenario selection model was implemented by the compromise programming method in MCAT software. Considering the budget limitation and available facilities on site, optimum allocations to recycling, composting, incineration and landfilling methods were obtained as 115,486, 132,094, 71,905 and 45,516 tons/year, respectively. The results of this study indicated that the metaheuristic algorithm in MCAT software was an efficient tool in decision making about waste management systems and thus, it was suggested to municipality managers and regional planning authorities.

Introduction

Municipal solid wastes are one of the most important environmental issues and have now become a global challenge [1, 2]. In developing countries, most of the produced waste is being dumped into non-engineered landfills [3]. In order to achieve an effective and sustainable waste management system, it is necessary to minimize waste production and maximize its recovery before it is disposed into a landfill. Based on the waste management hierarchy, different scenarios can be defined to divert waste from ending up into a landfill. Usually, an optimal waste management system requires a combination of methods. Various methods available for municipal waste management include composting, anaerobic digestion, waste incineration, recycling, and finally landfilling. Finding the best combination implicates uncertainty and multiple goals; thus it is a complex decision-making problem [4]. In this regard, effective scenarios should include proper waste processing methods that are selected based on the characteristic of waste as well as the cost and environmental performance of facilities [5]. A review of references on planning and management of municipal waste shows a growing trend in the number and complexity of relevant mathematical models. Several models focus solely on economic aspects and intend to minimize costs while neglecting the potentially adverse environmental impacts [6]. While cost is important in optimizing waste management strategies, it is also necessary to consider environmental issues [7]. Indeed, a sustainable waste management system accommodates economic and environmental issues together. Due to conflicting goals in multi-objective optimization models, a set of optimum solutions may apply instead of a single one. Therefore, different optimization algorithms based on different techniques have been developed to deal with the environmental problems of municipal solid waste management systems [8–10].

In this study, we considered costs along with environmental performance criteria to optimize waste management in the southwest region of Tehran province. A multi-criteria analysis approach was carried out by the Multi-Criteria Analysis Tool (MCAT). MCAT combines the Metaheuristic Algorithms with Multi-Criteria analysis to support decision-making efforts in environmental and natural resources management [11]. The model may include both constraints and criteria in selecting the best scenario. The main core of MCAT is structured by Compromise Programming Approach [12].

Some studies are focused on waste allocation. Najm et al. [4] and Rathi [13] utilized a linear programing (LP) model for the optimisation of waste allocation to some existing facilities. Few researchers have also incorporated uncertainty into waste allocation problems. Xu et al. [14] and Li, Chen [15] used the LP model with interval-parameters to carry out estimations under uncertainty. Moreover, Wu et al. [16] created a scenario-based fuzzy-stochastic quadratic programming model (SFQP). Others have considered environmental issues in waste allocation models. Chang et al. [17] developed a multi-objective linear programming model (MOLP) with an environmental objective function to minimise methane and carbon dioxide emissions. Similarly, Minoglou, Komilis [18] introduced a non-linear programing model (NLP) that enabled obtaining the least possible amount carbon dioxide emissions.

Other studies have considered the complex problem of allocation alongside with facility capacity expansion. Yedla, Sindhu [19]; Dai et al. [20] and Huang et al. [21] used an interval-parameter mixed integer linear programming model (MILP) and incorporated uncertainty into the complex problem of allocation and capacity expansion. Likewise, Mavrotas et al. [22] and Lyeme et al. [23] optimised the complex problem to determine facilities as well as their allocations and capacity expansions in Athens and Dar-e-Salaam respectively.

Facility and site selection problems are also observed among waste management researches. Münster, Meibom [24] developed an LP model to identify the optimum allocation of the waste-to-energy (WTE) facility. Badran, El-Haggar [25] introduced a MILP model for selecting municipal solid waste collection sites. Yadav et al. [26] introduced a systematic approach that featured an NLP model along with the geographical information system (GIS) to determine the best distribution of waste transfer stations.

Waste transfer vehicle routes also engage an optimization problem. Bonomo et al. [27] and Louati [28] determined the least travel distances by integer programing (IP) and MILP approaches. Arribas et al. [29] used IP with GIS tools to minimise time and the operation and transportation costs of waste collection.

The design and optimisation of an MSW recycling network is also an area of research. Santibañez-Aguilar et al. [30] developed a multi-objective mixed linear programming model (MOMILP) to optimize the types and locations of processing facilities as well as waste and resultant product distributions. Ng et al. [31] developed a WTE supply network for optimizing facility types, as well as their locations and optimal allocations. Similarly, Tan et al. [32] developed a MILP model to design a waste recycling network with respect to environmental issues, such to determine the best complex of waste treatment facilities, also their optimum capacities and waste allocations.

Waste management research has been carried out to develop different types of optimization models including LP, MILP, NLP, NSGA-II, MOMILP, SFQP, multi-objective programming and stochastic programming. Aside from problem-solving methodologies and models, the literature on research can be investigated in terms of practicability. Tilted towards economy, available research on waste management does not address the three pillars of sustainability evenly. The environmental pillar is poorly respected and social issues are neglected. Authors believe that waste management research lacks optimisation models that consider all three pillars of sustainability.

In the recent decade many studies have been carried out on solid waste management in Iran. Erfani et al. [33] developed a new GIS-based method to optimize waste collection routes and waste bin locations in Mashhad city. Results demonstrated that the existing system was incompatible with the urban structure and population distribution and thus, it required optimization. Ahani et al. [34] developed the non-dominated sorting genetic algorithm (NSGA-II) with two objectives: minimized environmental and economic costs to optimize waste management in Tehran city. Mofid-Nakhaee, Barzinpour [35] analysed a multi-compartment capacitated arc routing problem (CARP) with supplemental waste collecting facilities in Tehran city by developing adaptive large neighbourhood search (ALNS) and hybrid ALNS including whale optimization algorithms. Results indicated that multi-compartment vehicles (MCV) were more cost-effective than current operating vehicles. Babaee Tirkolaee et al. [36] designated the robust CARP for collecting municipal waste and designed a hybrid simulated annealing (SA) heuristic algorithm with an efficient cooling equation to yield appropriate results for the problem. The algorithm was proven efficient by comparing it with the mathematical model proposed for waste management in Esfahan city of Iran.

The study area has emerged as a fast-growing region. Despite its rapid growth of population and settlements, the majority of its municipal solid waste is disposed into a landfill that lacks an appropriate waste management system. No comprehensive plan to diversify waste management facilities in favour of environmental sustainability in near future is in place. According to the Waste Management Act [37], the Iranian municipalities are obligated to develop management plans and increase the waste recycling rates according to their local socioeconomic status, waste generation rate, waste composition and budget. The goal of this study is to carry out analysis to provide a better insight for using waste as a source of energy and materials while incorporating economic and environmental issues at the same time. The output of this research provides an optimal configuration of a waste management system in the study area through comparing three scenarios that best meet the economic-environmental concerns of waste industry managers of the study area.

This paper is organized as presenting the multi objective model development (including problem description, mathematical and optimization models and the proposed solution method) in "Materials and methods" section, results and discussion (including optimization results, sensitivity analysis and numerical results from the implementation of the algorithm) in "Results and discussion" section and finally, conclusion and directions for future research in the study area in "Conclusions" section.

Materials and methods

Study area

The study area consists of 12 cities and a number of villages in the southwestern region of Tehran province (Fig. 1). The region generates 365000 tons of waste per year which is equivalent to 1000 tons/day. All of the waste generated in the study area is transported to Rudeshur landfill after a limited manual sorting by illegal collectors. Rudeshur landfill site (35˚ 25’ 50” N, 50˚ 58’ 39” E) is located at 30 km southwest of Tehran city and in Robatkarim county. Table 1 shows the waste disposal amounts of different residential areas to the landfill in 2019.

Fig. 1.

Study area and landfill location

Table 1.

Residential areas served by Rudshur landfill and their waste quantity

	Residential area	Estimated population in 2019	Waste generation rate (ton/day)
1	Chardangeh	52,077	31
2	Vahidiyeh	34,601	23
3	Nasirabad	29,808	25
4	Nasimshahr	219,613	127
5	Golestan	266,885	161
6	Salehiyeh	62,275	43
7	Parand	97,292	109
8	Ahmadabad	22,462	15
9	Robt karim	118,439	80
10	Eslamshahr	479,298	274
11	Shahedshahr	26,583	15
12	Sabashahr	56,165	27
13	Villages	138,215	70
Total		1,303,713	1000

Multi Criteria Analysis Tool (MCAT)

To optimize solid waste management methods in the study area, MCAT was applied. This software was initially developed by the Commonwealth Scientific and Industrial Research Organization (CSIRO) in 2007 as a decision support system for environmental and natural resources management. Later, many studies applied this method with favourable outcomes [38–40]. The main structure of MCAT that is developed from the Multi-Criteria technique is based on the compromise programming approach. In compromise programming, a decision is selected according to its distance from a desired ideal condition (user-defined). The general form of the benefit score equation in the compromise approach is calculated using Eqs. 1, 2 and 3.

$$U_j^{-}={\left[{\sum }_{i=1}^M{W}_i^c{\left(1-\frac{f_i^{+}-f_{ji}}{f_i^{+}-f_i^{-}}\right)}^c\right]}^{\frac{1}{c}}$$ 1

Where Uj is the benefit score of decision j; M the number of criteria; $W_i^c$ the weight of criterion i; c the importance factor of maximal deviation from ideal solution; $f_i^{+}\text{and}{f}_i^{-}$ the highest and lowest scores, respectively; and finally, $f_{ji}$ represent the performance score of criterion i in decision j.

The simplified equation by the following linear transform will result:

$$g_{ij}=\frac{f_i^{+}-f_{ji}}{f_i^{+}-f_i^{-}}$$ 2

$$U_j^{-}={\left[{\sum }_{i=1}^M{W}_i^c{\left(1-g_{ij}\right)}^c\right]}^{1/c}$$ 3

In this case, gij can be replaced by a value between 0 and 1.

In reality however, environmental factors interact with each other and hence, environmental criteria cannot be described by linear functions. Therefore, the few non-linear transform relations of sigmoidal, convex and concave transforms are included in MCAT [41]. Convex Transform (Eq. 4), Concave Transform (Eq. 5), and Sigmoidal Transform (Eq. 6) are most common.

$$U_j^{-}={\left[{\sum }_{i=1}^M{W}_i^c{\left(1-\sqrt{1-(g_{ij}-1}{)}^2\right)}^c\right]}^{1/c}$$ 4

$$U_j^{-}={\left[{\sum }_{i=1}^M{W}_i^c{\left(\sqrt{1-{g_{ij}}^2}\right)}^c\right]}^{1/c}$$ 5

$$U_j^{-}={\left[{\sum }_{i=1}^M{W}_i^c{\left(e^{{(log}_e\left(0.5\right)g_{ij}^d)/m^d}\right)}^c\right]}^{1/c}$$ 6

Figure 2 shows the context of MCAT. Input to MCAT includes a budget constraint (1), a set of decision options (with corresponding costs) and a set of criteria (2). With the input, MCAT performs multi-criteria analysis (3) by using subsequent complex optimisation models that are developed from cost utility ratios of the decision options and indicate how effectively the expenditure is spent (4 and 5). The interfaces and charts of sensitivity analysis provide visual means to observe the impacts of changing boundary conditions. Ultimately, the results will support the decision making process.

Fig. 2.

The decision support workflow using MCAT

The appropriate transform relation is determined based on factors such as the target problem, data availability, and stakeholder benefit/cost ratios. Table 2 describes the selected transform functions in this research.

Table 2.

Transformation functions in MCAT

Criteria	Transform type	Description
Surface and groundwater pollution	Convex $U_j^{-}={\left[\sum _{i=1}^M{W}_i^c{\left(1-\sqrt{1-(g_{ij}-1}{)}^2\right)}^c\right]}^{1/c}$	The smallest value is most desirable and gives the highest benefit score; increase in water pollution causes a rapid decrease in the benefit score (smallest criterion value is most desirable)
Benefit/cost	Linear $U_j^{-}={\left[{\sum }_{i=1}^M{W}_i^c{\left(1-g_{ij}\right)}^c\right]}^{1/c}$	Benefit/cost are linearly transformed; benefit scores increase as a function of transformed value (largest criterion value is best)
Soil pollution	Convex $U_j^{-}={\left[\sum _{i=1}^M{W}_i^c{\left(1-\sqrt{1-(g_{ij}-1}{)}^2\right)}^c\right]}^{1/c}$	The smallest value is best; receives larger scores for smaller values; increasing soil pollution decreases benefit scores rapidly (smallest criterion value is most desirable)
Noise pollution	Convex $U_j^{-}={\left[\sum _{i=1}^M{W}_i^c{\left(1-\sqrt{1-(g_{ij}-1}{)}^2\right)}^c\right]}^{1/c}$	The smallest value is best; receives large scores for smaller values; increasing noise pollution decreases utility scores rapidly ( smallest criterion value is ) most desirable
Job creation	Linear $U_j^{-}={\left[\sum _{i=1}^M{W}_i^c{\left(1-g_{ij}\right)}^c\right]}^{1/c}$	Job creation has a linear transform; utility scores are a direct function of transformed value. (largest criterion value is most desirable)
Odor	Concave $U_j^{-}={\left[\sum _{i=1}^M{W}_i^c{\left(\sqrt{1-{g_{ij}}^2}\right)}^c\right]}^{1/c}$	Effect of odor on waste management optimization changes non-linearly. (smallest criterion value is most desirable)
Landscape	Linear $U_j^{-}={\left[\sum _{i=1}^M{W}_i^c{\left(1-g_{ij}\right)}^c\right]}^{1/c}$	The landscape is linearly transformed; utility scores increase as a function of transformed value. (the smallest criterion value is most desirable)
Air pollution	Convex $U_j^{-}={\left[\sum _{i=1}^M{W}_i^c{\left(1-\sqrt{1-(g_{ij}-1}{)}^2\right)}^c\right]}^{1/c}$	The smallest value is most desired; receives a high score with low values; air pollution increase sharply reduces the utility score(smallest criterion value is most desirable)
Fauna and flora	Linear $U_j^{-}={\left[\sum _{i=1}^M{W}_i^c{\left(1-g_{ij}\right)}^c\right]}^{1/c}$	Fauna and flora have a linear transform; utility scores are a direct function of the transform result. (largest criterion value is most desirable)

In compromise programming, the user’s definitions are applied to determine the overall benefit score of each decision option. Also constraints such as budget applied as Boolean factors in decision making are considered. The knapsack problem is a good example of the application of constraint factors, in which a decision option is discarded despite scoring high on its overall benefit once its cost exceeds the budget [38]. MCAT finds the optimum solution combination via the branch and bound method [42]. In this method, all solution options are sorted from greatest to smallest in terms of their benefit/cost ratio. All possible solutions form a rooted tree that its branches are made of solution sets. Prior to selecting the optimal solution branch, its upper and lower bounds are estimated. If these values fail to produce a solution better than the one selected previously, that branch is ignored. The algorithm continues until it finds a branch that provides a solution that is equally or more appropriate than the bounding values of all branch nodes [43].

The Meta Heuristics methods of Hill-Climbing Local Search and Tabu Search (TS) are applied in MCAT for solution finding. In this study, Hill Climbing was used for the knapsack problem. Also, all decision options are arranged from high to low based on their benefit over cost ratio. Local searching algorithms go through the sorted list, from one solution to another checking the ones nearby at the same time. At first, the decisions with the highest benefit/cost ratio are considered, among which the final decision is ultimately determined based on total cost values [44, 45].

Although more time-consuming than Hill Climbing, TS is applied to ensure a more reliable overall local searching performance. If other search methods are missing, TS may continue to search branches even if no improvement is obtained. TS method identifies the initial solution randomly with regard to constraints.

Scenarios for optimizing waste management methods

Several scenarios for optimizing solid waste management were developed according to environmental and socioeconomic goals, three of which were explicitly selected by environmental and economic experts through the Analytical Hierarchy Process (AHP) [46]. AHP is a three-step technique that performs pairwise comparisons for the criteria [47]. In the first step called decomposition, a complex decision is organized into simply manageable elements. The second step, comparative judgment, associates with evaluating priorities in pairwise comparisons by asking expert choice. Finally, the synthesis of priorities is carried out in the third stage to identify the consistency ratio of criteria. Consistency ratio is used to assess judgments and it features a value between zero and one and if not, the judgments would require reassessment [48, 49].

Economic cost

The total net cost (total costs minus revenues) of implementing waste management facilities (recycling, composting, incinerator, and landfilling) is regarded. Rudshur landfill was considered a conventional landfill without a gas collection system, so no revenue was obtained. Total costs of each facility are estimated from its construction costs (CC), investment return rate (RR), and operation costs (OC) and input waste rate (IWR) in ton/year using Eq. 7.

$$Totalcostfacilityi=\frac{(CC\times RR)+OC)}{IWR}$$ 7

The return rate of investment is calculated based on the real interest rate (IR) over a year and estimated life span (ELS) of the facility using Eq. 8.

$$RP=\frac{IR\times {\left(1+IR\right)}^{ELS}}{{(1+IR)}^{ELS}-1}$$ 8

The cost of each management method and other relevant data (regularly collected) were obtained from the database of Robatkarim Waste Management Organization (Table 3). Also, the revenue of facility i (Revenue i) is calculated based on the waste input to facility ($Xi$), the percentage of component (j) of waste to facility ($Pij$), and revenue from component j of the input waste ($PCij$) as described in Eq. 9.

$$Revenuei={\sum }_{j=1}Xi\times Pij\times PCij$$ 9

Table 3.

Economic cost of the municipal waste management facilities

Facility		Cost			Revenue
		Annual cost		Unit (Us$/Ton)	Annual revenue		Unit (Us$/Ton|)
Recycling		1023.41		8.845	1263.4		10.92
Compost		1762.95		6.900	1072.07		4.196
Incinerator		3823.7		104.759	3600		8.6569
Landfill		844.7		5.786	-		-

Results and discussion

Analysis of quantity and quality of waste produced

The amount of waste and its composition determine the appropriate facilities for sustainable waste management affect the results of feasibility assessment of material and energy recovery and finally, specify the design of necessary equipment. Around 68 % of the waste transported to the landfill was made of organic matter, whereas 29.5 % was composed of recyclable materials. Table 4 shows the amount and composition of waste transported to the landfill in 2019.

Table 4.

Physical components of waste in the study area

Physical components	Percentage of components	Annual waste amount (ton )
Paper	7.1	25,915
Cardboard box	6.3	22,995
Plastic Bottles	1.4	5110
Plastic and Nylon	3.7	13,505
Metal	2.8	10,220
Glass	2.1	7665
Bread	5.5	20,075
Textiles	0.6	2190
Others	2.2	8030
Organic waste	68.3	249,295

Environmental impacts of different waste management methods

The adverse environmental impacts of different municipal solid waste management options are shown in Table 5. Impacts depending on their nature and severity may range from global, regional, and local disturbances. Some impacts include air pollution, greenhouse gas (GHG) emissions, noise, odours, soil and water contamination, and visual pollution; and may lead to secondary impacts such as climate change, health threat, loss of species, degradation of land, etc. The severity and significance of impacts depend on factors of distance from the landfill, waste composition, site environmental condition, dominant wind direction, disposal method, and the age of site [50].

Table 5.

Significant environmental criteria for assessment of municipal waste management options

Activity	Noise	Odor	Fauna and flora	Soil	Water quality	Air quality	Weather	Dust
Recycling of materials	x	x	x	x	xx	xx	-	x
Aerobic composting	xx	xxx	-	-	xx	xxx	x	xx
Incineration	xx	xx	xxx	xxx	xx	xxx	x	xxx
Landfill	xxx	xxx	xxx	xxx	xxx	xxx	xxxx	xx

Optimization results

The evaluation matrix is the main core of MCAT by which all of the different solutions, waste management methods, criteria and properties are analysed (Table 6). In this matrix, criteria are indicated in columns, whereas the rows correspond to solution options.

Table 6.

Evaluation matrix for solid waste management method optimization

Criteria option	Surface water pollution	Ground water pollution	Soil pollution	Air pollution	Landscape	Fauna and flora	Noise pollution	Odor	Benefit and cost	Job creation
Recycling	0.173	0.032	0.168	0.058	0.077	0.036	0.033	0.101	0.141	0.181
Incinerator	0.132	0.046	0.095	0.252	0.086	0.059	0.087	0.164	0.045	0.034
Compost	0.143	0.043	0.075	0.075	0.0812	0.047	0.036	0.211	0.153	0.157
Landfill	0.202	0.709	0.204	0.059	0.119	0.047	0.038	0.173	0.04	0.040

The weights, as shown in Table 7, show the relative significance of each criterion that was defined by expert choices. The efficiency of AHP is assessed by consistency ratio (CR) efficiency that fell into the acceptable limit (< 0.10) in this study indicating that the weights were valid. Consequently, three scenarios were obtained: Environmental scenario (scenario1), Economic scenario (scenario2), and Environmental- Economic scenario (scenario3).

Table 7.

Criteria weights

Criteria Weights	Surface water pollution	Groundwater pollution	Soil pollution	Air pollution	Landscape	Fauna and flora	Noise pollution	Odor	Benefit and cost	Job creation
S1	22	6	19	14	7	5	5	16	3	3
S2	12	3	11	8	4	3	3	9	24	23
S3	17	4	15	11	6	5	3	12	14	13

Table 8 shows the optimization results for the selected solid waste management scenarios in terms of minimized environmental pollution with maximized socio-economic returns.

Table 8.

Result of MCAT optimization between competing waste management methods

Options	Benefit			Funded			Rank			Rel.allocation
	S1	S2	S3	S1	S2	S3	S1	S2	S3	S1	S2	S3
Recycling	0.739	0.93	0.787	Yes	Yes	Yes	2	1	2	27.8	35.84	31.64
Incinerator	0.62	0.35	0.49	Yes	NO	Yes	3	4	3	23.31	13.49	19.7
Compost	0.87	0.84	0.9	Yes	Yes	Yes	1	2	1	32.72	32.55	36.19
Landfill	0.43	0.47	0.31	NO	Yes	NO	4	3	4	16.17	18.11	12.47

The result of solid waste management optimization for scenario1 indicates that compost method ranked first (with benefit/cost ratio of 0.87) and processing and recycling ranked second (with benefit/cost ratio of 0.73). Finally, the rank of landfill method (with benefit/cost ratio of 0.43) was last and so not recommended in the study area.

The result of solid waste management optimization for scenario2 indicated that processing and recycling methods were most profitable whereas, incineration was least. The results of scenario3 ranked compost first, recycling second and landfill last. Interestingly, landfill did not meet the economic and environmental objectives in the study area according to both scenarios 1 and 3.

The shares of total benefits for different waste management methods are indicated in Fig. 3. Accordingly, in scenario 2 the recycling method yielded the highest benefit percentage (37 %) of the total economic benefit while waste incineration held the lowest. In scenario1, landfill had a lower share (13 %) of total benefits, compared to other methods. Finally, integrating the overall benefits in scenario3, composting ranked highest.

Fig. 3.

Sensitivity analysis for boundary conditions

Sensitivity analysis of waste management methods against different criteria

In this study, MCAT performed sensitivity analysis by assigning a weight between 0 and 100 to each criterion at first. Then, as illustrated in Fig. 4, it distributed the difference between that weight value and 100 equivalently among other criteria. In the next step and repetitively, MCAT recalculated the benefit scores with increasing the weight of each criterion individually and reducing others, until the criterion weight reached 100. The total sum of weights were kept constant (100) at all times. Finally, benefit score values were plotted against criterion weight values for each disposal method. In this illustration, a graph with a steeper slope indicated higher sensitivity of that waste management method to that particular criterion. Consequently, landfilling compared to other methods, had a steep descending graph for the majority of criteria. For the criterion of job creation, compost and recycling had a steep ascending slope and thus they were highly sensitive to weight changes. Reversely, the waste incinerator method had a steep descending slope for the same criterion. For the criterion of benefit to cost, landfill and waste incinerator methods had a sharp decreasing trend whereas, compost and recycling methods were both increasing for all criteria.

Fig. 4.

Sensitivity analysis of different disposal methods for each criterion: (a) criterion air pollution, (b) criterion odour, (c) criterion employment, (d) criterion surface water pollution, (e) criterion benefits, (f) criterion animal, (g) criterion soil pollution, (h) criterion noise pollution, (i) criterion groundwater pollution, and (j) criterion landscape

Comparison of the percentages of waste allocations in different scenario

Tables 9, 10, 11 and 12 show the optimum annual waste allocations and their relative capacities for different disposal methods. Based on study the objectives and their obtained results, scenario3 in which economic costs were minimized with increasing revenue and reducing adverse environmental impacts was selected. The optimum capacities of recycling, composting, incineration, and landfill methods yielded 115,486 tons/year, 132,094 tons/year, 71,905 tons/year, and 45,516 tons/year respectively. In Fig. 5, waste allocations are compared in scenarios 1, 2 and 3.

Table 9.

The optimum design capacity and waste allocation to the recycling method

Scenario	The optimum allocated flow of waste (ton /year)	Optimum allocation (%)
1	101,473	27.8
2	130,816	35.84
3	115,486	31.64

Table 10.

The optimum design capacity and waste allocation to the aerobic composting method

Scenario	The optimum allocated flow of waste (ton /year)	Optimum allocation (%)
1	85,045	23.3
2	48,910	13.4
3	71,905	19.7

Table 11.

The optimum design capacity and waste allocation to the incineration method

Scenario	The optimum allocated flow of waste (ton /year)	Optimum allocation (%)
1	85,045	23.3
2	48,910	13.4
3	71,905	19.7

Table 12.

The optimum design capacity and waste allocation to the landfilling method

Scenario	The optimum allocated flow of waste (ton /year)	Optimum allocation (%)
1	59,021	16.17
2	66,102	18.11
3	45,516	12.47

Fig. 5.

Comparison of the percentage allocated to the facilities for different scenarios

Discussion

Municipal waste management of cities in Tehran province suffer from high economic and environmental costs with a worsening trend due to population growth and lacking appropriate management systems. Economic aspects and environmental performances appear as conflicting objectives in optimization of waste management in the study area.

Among all methods, landfill ranked last in scenario1 because of its adverse environmental impacts the most significant of which are air pollution, contamination of water resources and proximity to Shur river. Landfill ranked third in scenario 2 because of its low cost-benefit rate. Since landfill requires less specialized equipment, facilities, and manpower, comparing to other disposal methods, it is one of the cheapest and simplest methods of municipal waste management. Despite environmental pollution and land scarcity accompanied by the high cost of land in Tehran province, this method has become most popular in the region. Yet, this method was rejected in all three scenarios (including the economic scenario). In scenario2, recycling ranked one because of its high profitability. Although processing and recycling of waste have advantages such as reducing landfill input, valuable waste extraction, job creation for local people, high revenue, and finally low initial investment, it has not been practiced in the study area for various reasons. Among these reasons are lack of government support for source waste-separation plans, conversion industries, product sale markets, and inappropriate organization of applicants and contractors.

Compost ranked first in scenario3 because the great organic content of waste composition makes this type of waste valuable for agricultural practices. Therefore, by the use of modern processing methods along with marketing and creating awareness in the agricultural sector, its adverse environmental impacts can be mitigated and significant revenue may be achieved. Unfortunately, this method is neglected in present practice because necessary equipment is unavailable and there is no demand for compost fertilizer by farmers. Moreover, source separation of dry waste is not practiced, which increases the likelihood of compost fertilizer contamination.

The incineration method ranked last in scenario 2 and thus, it was rejected. This method associates with expensive initial implementation, high repair and maintenance costs, lack of policy to guarantee the purchase of produced electricity and finally, high water consumption.

Mavrotas et al. [22] resulted that the contributions of recycling and waste to energy methods were significant. Moreover, in scenario 1 (private cost only) which was the economic criteria, incineration was strongly determined as the preferred option whereas, in scenario 2 (private plus external costs) which included the environmental impacts along with costs, less incineration was desirable due to its external costs. Agreeably in our research, the incineration method was found less desirable when only environmental impacts were considered (scenario 1).

Yedla, Sindhu [19] employed an integrated approach and considered several waste disposal methods through multivariate functional models to reduce GHG emissions from municipal solid waste that was most economically beneficial. Consequently, aerobic composting was the optimal method. Our study proposes a combination of methods to minimize environmental costs (scenario 1). Also, investigating the trends of facility allocations, composting ranked highest in scenario 3 among all scenarios. Consequently, composting was the optimal method in terms of environment and economy together for managing municipal solid waste in the study area.

Deng et al. [10] illustrated that among all waste management methods the results of contribution analysis were very different for landfills and incineration facilities in the two scenarios of minimum cost and minimum GHG emissions. According to our study, in minimization of environmental and economic costs scenario, recycling and landfill methods were mostly different. The trend of recycling against economic and environmental costs indicated its economic preferability with minor impacts on the environment. For landfill facility, minor differences were observed in the environmental and economic scenarios and trend analysis showed that this facility is more desirable from an economic point of view. Huang et al. [21] indicated that the scenario in which three potential waste disposal methods (anaerobic digestion, incineration, and landfilling) were considered interactively, the best overall results were obtained in terms of GHG emissions. Our study reveals that to maintain minimum environmental costs (scenario 1), it is necessary to apply a combination of disposal methods that consist of composting, recycling, incineration and landfill. A hybrid system with all methods would minimize the overall costs.

The multi-criteria model proposed in this study enables regional planners to design optimized waste management methods to cut down environmental pollution and economic costs altogether, and make a leap towards sustainable development in the study area.

Conclusions

Problems of municipal waste management often involve optimization among several economic, environmental, technical, and social goals. Considering various constraints, decision-makers seek to find the most desirable and practical solution by analysing several scenarios. In this study, three scenarios were designed with focus on economy and the environment for optimizing the waste management practice in the southwest region of Tehran province. A team of environmental and economic experts and individuals engaged in management positions weighted the scenarios through the AHP. MCAT using compromise programming was applied for multi-criteria conflict resolution.

We applied budget and facility constraints in the optimization process, and based on the results of scenario 3, landfill was rejected for the study area. Similar analysis for scenario 2 did not recommend incinerators. The overall results of this study recommended a combination of methods to minimize the overall costs of environment and economy. The recycling and composting methods ranked highest because their associated costs were lowest in all scenarios. Also, scenario 3 was advantageous among the selected scenarios, because of reducing economic and environmental costs at the same time.

Comparing the results with existing condition, the current development plans that target facilities and equipment in the study area need to be changed. It is necessary to raise the level of awareness of managers and decision-makers to alter existing practices and current plans and pass new regulations and appropriate waste management policies. The results of this study can provide a basis for the optimization of other landfills in Iran and help develop relevant strategies and policies at regional and national scales.

Further research can target uncertainty of input data such as waste amounts produced at different residential areas by stochastic programming and dynamic modelling. Another area of research can be to insert waste collection vehicle routing decisions into the model. Smart optimization models for waste management methods can be considered in the overall problem optimization model. Taking note of the social aspect of waste management practices, factors as important it is recommended to highlight social factors as objective functions of optimization models in future research.

Data availability

Not applicable.

Code availability

Not applicable.

Declarations

Conflict of interest

The authors confirm no conflicts of interest associated with this publication.

Ethics approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

Not applicable.