An economic evaluation and assessment of environmental impact of the municipal solid waste management system for Taichung City in Taiwan

Abstract

Municipal solid waste management (MSWM) is an important environmental challenge and subject in urban planning. For sustainable MSWM strategies, the critical management factors to be considered include not only economic efficiency of MSW treatment but also life-cycle assessment of the environmental impact. This paper employed linear programming technique to establish optimal MSWM strategies considering economic efficiency and the air pollutant emissions during the life cycle of a MSWM system, and investigated the correlations between the economical optimization and pollutant emissions. A case study based on real-world MSW operating parameters in Taichung City is also presented. The results showed that the costs, benefits, streams of MSW, and throughputs of incinerators and landfills will be affected if pollution emission reductions are implemented in the MSWM strategies. In addition, the quantity of particulate matter is the best pollutant indicator for the MSWM system performance of emission reduction. In particular, this model will assist the decision maker in drawing up a friendly MSWM strategy for Taichung City in Taiwan.

Implications:

Recently, life-cycle assessments of municipal solid waste management (MSWM) strategies have been given more considerations. However, what seems to be lacking is the consideration of economic factors and environmental impacts simultaneously. This work analyzed real-world data to establish optimal MSWM strategies considering economic efficiency and the air pollutant emissions during the life cycle of the MSWM system. The results indicated that the consideration of environmental impacts will affect the costs, benefits, streams of MSW, and throughputs of incinerators and landfills. This work is relevant to public discussion and may establish useful guidelines for the MSWM policies.

Introduction

Municipal solid waste management (MSWM) is an important environmental challenge and subject in urban planning. A general life cycle of treating municipal solid waste (MSW) includes household collection, transferring, treatment, resource recycling, composting, and disposal. For sustainable MSWM strategies, the critical management factors to be considered include not only economic efficiency of MSW treatment but also its life-cycle assessment (LCA) of environmental impact. Some of the past studies have focused on mathematical programming (MP) techniques to be applied in the MSWM system, such as linear programming (LP) to achieve least-cost strategies with suitable constraint sets (Abou Najm and El-Fadel, 2004; Chang and Chang, 1998; Harrison et al., 2001; Lin et al., 2006; Wang et al., 2008), and mixed integer linear programming (MILP) to assess a feasible location and the capacity of new facilities in least-cost MSWM problems (Badran and El-Haggar, 2006; Chang and Lin, 1997; Chang et al., 2005; Xanthopoulos and Iakovou, 2009). Other programming techniques such as interval-parameter programming (Huang et al., 1992; Liu et al., 2009; Maqsood et al., 2004; Sun and Huang, 2010; Xu et al., 2009), probabilistic programming (Liu et al., 2009), fuzzy programming (Liu et al., 2009; Sun and Huang, 2010; Wang et al., 2011), stochastic programming (Wang et al., 2011; Xu et al., 2009), and quadratic programming (Li et al., 2011) have also been integrated into LP or MILP to obtain optimal strategies of MSWM whose parameters encounter uncertainties. Furthermore, some researches have analyzed solid waste management strategies with environmental impact limitations such as air pollution and leachate impacts (Chang and Wang, 1996; Chang et al., 1996a, 1996b), and noise control and traffic congestion limitations (Chang et al., 1996b). However, further investigations about the trade-off between the economical optimization and environmental impacts should be necessary.

LCA methodology is a widely accepted method for evaluating the environmental impacts associated throughout a product's life from raw material acquisition to production, use, and disposal (i.e., from cradle to grave). In the definition of LCA, the term “product” includes not only product systems but also service systems, such as MSWM systems (International Organization for Standardization [ISO], 1997). Therefore, LCA has been applied to compare the pollutant emissions from different MSWM strategies (Aye and Widjaya, 2006; Eriksson et al., 2005; Finnveden et al., 2005; Mendes et al., 2004; Mohareb et al., 2008; Thorneloe et al., 2007; Zhao et al., 2009).

Up to now, the previous MSWM studies only evaluated individual aspects of economical optimization (Abou Najm and El-Fadel, 2004; Badran and El-Haggar, 2006; Chang and Chang, 1998; Chang and Lin, 1997; Chang et al., 2005; Harrison et al., 2001; Huang et al., 1992; Li et al., 2011; Lin et al., 2006; Liu et al., 2009; Maqsood et al., 2004; Sun and Huang, 2010; Wang et al., 2008, 2011; Xanthopoulos and Iakovou, 2009; Xu et al., 2009) and environmental impact (Aye and Widjaya, 2006; Eriksson et al., 2005; Finnveden et al., 2005; Mendes et al., 2004; Mohareb et al., 2008; Thorneloe et al., 2007; Zhao et al., 2009). The integration of economic factors and environmental impacts is seldom discussed. In this paper, an optimization model, which integrates aspects of both economic and environmental impact, is applied to investigate the correlation between economical optimization and air pollution emissions for the life cycle of MSWM systems. The air pollutants considered in this study were sulfur oxides (SO_x), nitrogen oxides (NO_x), carbon monoxide (CO), and particulate matters (PM). The LP-based optimization model was used to determine the least-cost MSWM strategies with different emission reduction requirements of the air pollutants. A case study based on real-world operating parameters in Taichung City is also presented.

Case Study

Taichung City is located in west-central Taiwan. It has a total area of 2214.90 km² with a population of approximately 2.63 million. The area is composed of 29 administrative districts, and contains 22 MSW collection stations, three incinerators, six compost facilities, five landfills, and one bottom ash reuse facility, as shown in Figure 1. In Taiwan, the household collection of MSW is classified as general solid waste (GSW), recyclable materials, and kitchen waste. Currently, the daily MSW generation in Taichung City is about 3011 ton/day, of which GSW, recyclable materials and kitchen waste is about 72%, 22%, and 6%, respectively (Environmental Protection Administration [EPA] of Taiwan, 2009).

Figure 1. Geographical distribution of districts, collection stations, incinerators, compost facilities, and landfills in Taichung City.

Figure 2 shows the framework of the life cycle of MSWM in Taichung City. All MSW must be sent to collection stations. Then the recyclable materials are sorted at the collection stations and sold. The GSW is transferred to incinerators, and the kitchen waste can be sold for feeding swine or transferred to compost facilities. Rejected materials after resource recycling and composting must be transferred to incinerators, and residuals of reused bottom ashes must be transferred to landfills. The incineration bottom ashes are transferred to landfills or bottom ash reuse facilities; but the fly ashes can only be transferred to landfills.

Figure 2. The entire life cycle of MSWM system currently operated in Taichung City.

Although economic factors are very important concerns when developing a MSWM system, the environmental impacts should also be considered. The air pollution emissions are one of the major environmental impacts during the life cycle of MSWM systems. Most of the air pollutants are emitted during MSW transportation and incineration. Since the real-world emission data available in Taiwan EPA database for LCA investigation are only for CO, NO_x, SO_x and PM, this study employs these four pollutants as environmental impact indicators of MSWM systems. The variation of the least-cost MSWM strategies and unit costs are also examined when various pollution emission reduction requirements are applied.

Optimization Model

The optimization model developed in this study aims to obtain least-cost MSWM strategies covering the management, household collection, transfer, treatment, resource recycling, kitchen waste for composting or feeding swine, and disposal of MSW. The framework of the optimization model is shown in Figure 2. The constraints consist of mass balance requirements, capacity limitation of facilities, and air pollution emission limitations.

The objective is to minimize the total net cost of the MSWM system, which is the total transportation costs plus the treatment cost minus the total revenue from sales of electricity, recyclable materials and kitchen waste for feeding swine or compost, as shown in eq 1:

(1)

where TRAN_COST is the total transportation cost including the collection and transferring cost (NT$/day); TREAT_COST is the total treatment costs (NT$/day); and REVE is the total revenue (NT$/day). The total transportation cost TRAN_COST can be expressed as

where CC and TC are the unit collection cost and unit transferring cost (NT$/ton-km), respectively; D_xy is the distance from x (x = i, k, j, f, e) to y (y = k, j, f, l, e) (km); GSW_ik is the amount of GSW collected from district i and sent to collection station k (ton/day); REC_ik is the amount of recyclable materials collected from district i and sent to collection station k (ton/day); KW_ik is the amount of kitchen waste collected from district i and sent to collection station k (ton/day); MIXSW_kj is the amount of GSW mixed with the rejected materials sorted from recyclable materials from collection station k and transferred to incinerator j (ton/day); KW_kf is the amount of kitchen waste transferred from collection station k to compost facility f (ton/day); FASH_jl is the amount of fly ash transferred from incinerator j to landfill l (ton/day); BASH_jl and BASH_je are the amount of bottom ash transferred from incinerator j to landfill l and reuse facilities e (ton/day), respectively; RESFW_fj is the amount of residual materials from compost facility f transferred to incinerator j (ton/day); RESBASHR_el is amount of the residual materials transferred from bottom ash reuse facility e to landfill l (ton/day).

The total treatment costs TREAT _COST can be expressed as

where RR_ir is the ratio of category r material to all recyclable materials of district i; RRC_r is the unit treatment cost of material r (NT$/ton); SFD_k (ton/day) and SFDC_k (NT$/ton) are respectively the amount and unit treatment cost of kitchen waste for feeding swine of collection station k; TC_j is the unit treatment cost of incinerator j (NT$/ton); KWCC_f is the unit treatment cost for kitchen waste at compost facility f (NT$/ton); FASHC_l and BASHC_l are the unit treatment costs for fly ash and bottom ash of landfill l (NT$/ton), respectively; and BASHC_e is the unit treatment cost for bottom ash of reuse facility e (NT$/ton).

The total revenue REVE is

(4)

where RRSP_r is the selling price of recyclable material r (NT$/ton); SFDSP_k is the selling price of kitchen waste for swine feeding (NT$/ton); WETC_j , EP_j , and SR_j are respectively the waste-to-electricity transfer coefficient (WETC) (kWhr/ton), the selling price of electricity (NT$/kWhr), and the proportion of electricity generated by incinerator j which is sold; COMPGR_f is the compost generation rate of compost facility f (%); and COMPSP_f is the selling price of compost of compost facility f (NT$/ton).

The GSW, recyclable materials, and kitchen waste generated by each district should be shipped to available collection stations, as reflected in eqs 5, 6, and 7. At each collection station, the recyclable materials can be sold, and the sorted rejected materials mixed with GSW should be transferred to incinerators, as reflected in eq 8. The kitchen waste can be sold for feeding swine or transferred to available compost facilities, as reflected in eq 9. For each compost facility, compost residuals should be transferred to incinerators, as reflected in eq 10. For each incinerator, the bottom ash can be transferred to available landfills or bottom ash reuse facilities, whereas the fly ash can only be transferred to landfills, as reflected in eqs 11 and 12. The residue of reused bottom ash needs to be transferred to landfills, as reflected in eq 13.

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

where GENE_i is the MSW generated in district i; α _i and β _i are respectively the ratios of recyclable materials and kitchen waste to MSW in district i; RESKWGR_f is generation rate of residuals for compost; FASHGR_j and BASHGR_j are respectively the generation rates of fly ash and bottom ash in incinerators j; and RESBASHRR_e is the generation rate of residuals for bottom ash reused.

The designated treatment capacities for each collection station, incinerator, compost facility and landfill should not be exceeded, as expressed in eqs 14–18.

(14)

(15)

(16)

(17)

(18)

where CAP_k , CAP_j , CAP_f , CAP_l , and CAP_e are respectively the designed capacities (ton/day) for collection station k, incinerator j, compost facility f, landfill l, and bottom ash reuse facility e.

For a MSWM system, the total emission of each air pollutant was calculated using the following formula:

(19)

where EM_z is the total emissions of pollutant z (z = CO, NO_x, SO_x, or PM) for MSWM system (kg/day); EMC_z and EMT_z are the emission factors of pollutant z for collecting and transporting MSW (kg/km), respectively; CAR_C and CAR_T are the capacities of collection vehicles and transportation vehicles (ton/vehicle), respectively; EMF_INC_jz , EMF_KW_fz , and EMF_SFD_kz are the emission factors of pollutant z for incineration, composting, and feeding swine (kg/ton), respectively.

Description of the Parameters

There are several physical, economic, and environmental parameters required by this study for each component of the MSWM system. The amounts of MSW and its composition generated by each administrative district are available from the Taiwan Environment Data Warehouse (EPA of Taiwan, 2009). The air pollutant emission coefficients of vehicles transporting MSW were obtained from the Taiwan Emission Data System version 7.0. The emission coefficients of NO_x, SO_x, CO, and PM for collection are 16.17, 0.0322, 8.45, and 1.2513 g/km, respectively, and those for transferring are 13.86, 0.0283, 4.23, and 1.2513 g/km, respectively (EPA of Taiwan, 2007). Recycling of kitchen waste and other recyclable materials is implemented at each collection station, which the treatment cost and income from recyclable materials is shown in Table 1. These data come from two investigative EPA reports of resource recycling in Taiwan. The operation parameters of three incinerators can be obtained from Taiwan Environment Data Warehouse of EPA of Taiwan, as shown in Table 2. The parameters required for environmental emission factors are regarding to continuous monitor air pollution for stack on incinerators, as shown in Table 2. The six compost facilities are respectively Taichung City, Wufong, Fengyuan, Shihgang, Sinshe, and Waipu. The design capacities are 10, 5, 8, 4, 0.3, and 100 ton/day, respectively (EPA of Taiwan, 2004, 2009). In addition, the unit treatment cost, unit benefit, and coefficient in which kitchen waste converted into fertilizer are 575 NT$/ton, 1500 NT$/ton, and 0.2 ton/ton, respectively (EPA of Taiwan, 2004, 2008b, 2009). Only air emission of NO_x and SO_x are considered for the compost, whose emission factors are 0.00129 and 0.176 kg/ton, respectively (Tsai, 2005). The capacities of the five landfills in Taichung City, Shengang, Houli, Dali, and Wufong are 180, 55, 62, 150, and 80 ton/day, respectively. In the following, this study will use above parameters to estimate various MSWM strategies for Taichung City in Taiwan.

Table 1. The treatment costs and selling prices of recyclable materials

Term	TreatmentCost(NT$/kg)	Selling Price(NT$/kg)
Paper (EPA of Taiwan, 2007)	0.54	2.60
Iron	10.50	7.80
Aluminum	4.10	60.50
Other metal	10.50	7.80
PET	14.33	21.60
Plastic	2.62	10.60
Glass	6.90	0.80
Electrical appliance	7.82	3.22
Tire	7.20	5.00
Teltra pak container	3.11	2.60
Paper container	0.54	2.60
Battery	40.00	15.00
Lead-acid battery	3.28	50.60
Computer	25.04	6.32
Light bulb	6.24	2.50
Other plastic	25.00	3.00
Agriculture agent container	1.00	4.50
Kitchen wastes for swine feed (EPA of Taiwan, 2008a)	1.00	1.40

Table 2. The operation parameters and pollutant emission coefficients of incinerators in Taichung City (EPA of Taiwan, 2009)

Operation Parameters/Emission Coefficients	Incinerator
	Nantun	Houli	Wurih
Operation parameters
Tipping fee (NT$/ton)	489	339	129
Generation rate of bottom ash (%)	14	17	15
Generation rate of fly ash (%)	5	5	6
WETC (kWhr/ton)	404	573	572
Rate of generated electricity sold (%)	75	82	80
Selling price of electricity  (NT$/kWhr)	1.74	1.68	1.83
Capacity (ton/day)	900	900	900
Pollutant emission coefficients
NOx (kg/ton)	1.44	1.45	1.23
SOx (kg/ton)	0.06	0.12	0.05
CO (kg/ton)	0.06	0.07	0.26
PM (kg/ton)	0.30	2.03	3.94

Scenario Study

According to the current circumstances in Taichung City, this study proposed two scenarios to assess the impacts of economic and environmental concerns of the MSWM strategies. The scenario of each alternative is described below:

Scenario 1: Scenario 1 is to find the least-cost MSWM strategy based on the existing facilities and operating parameters described above. The environmental impacts under such a least-cost strategy are also investigated and evaluated.

Scenario 2: Scenario 2 is the same as Scenario 1, except the emission reduction of the four air pollutants are individually required. The additional constraint can be expressed as eq 20:

(20)

where EM_BASE_z is the emission of pollutant z (z = CO, NO_x, SO_x, or PM) of Scenario 1 (kg/day); and RED_ratio_z is emission reduction ratio of pollutant z.

Results and Discussion

The optimization models developed in this study were solved by PuLP package Version 1.4.7 of PYTHON Version 2.6. The models consist of 2585 decision variables, 2138 parameters, and 299 constraints. The results obtained from the two scenarios are summarized in the following sections.

Results of Scenario 1

The results of the least-cost strategy in Scenario 1, including costs, revenue, and pollution emissions, are summarized in Table 3. It shows that the minimum net cost of the MSWM system in Taichung City is −712,509 NT$/day, indicating a profit of NT$712,509 per day. It was also found that the treatment of MSW is the most costly process in the MSWM system, especially the tipping fees for resource recycling. Most of the collection and transferring costs are used for GSW. The revenue is mainly achieved from the sale of recyclable materials, electricity, and kitchen waste for feeding swine. The income from recyclable materials of 3,178,418 NT$/day is the major revenue of the MSWM system, and results in a net profit of 951,653 NT$/day when the tipping fees of treatment for recyclable materials is covered. Although the sale of electricity from incinerators of 1,678,809 NT$/day is another major source of revenue for the MSWM system, its treatment and disposal costs require 1,634,349 NT$/day. Therefore, there is no significant net profit for incinerators.

Table 3. The least-cost results of MSWM strategy and pollution emissions for Scenario 1

Net cost (NT$/day)				−712,509
Total cost (NT$/day)				4,403,452
The collection cost (NT$/day)				174,356
The transportation cost (NT$/day)				183,172
The treatment cost (NT$/day)				3,031,469
The treatment cost of resource recycling (NT$/day)				2,226,765
The treatment cost of kitchen waste (NT$/day)				184,810
The treatment cost of general waste at the incinerators (NT$/day)				619,894
The disposal cost (NT$/day)				1,014,455
Total income (NT$/day)				5,115,961
The income of resource recycling (NT$/day)				3,178,418
The income of swine feed (NT$/day)				258,734
The income of electricity (NT$/day)				1,678,809
The emission amount of pollutants (kg/day)
Process	SOx	NOx	CO	PM
Transportation	0.98 (0.55%)	484.07 (14.45%)	198.79 (38.25%)	40.67 (0.73%)
Swine feed	0.06 (0.03%)	0.12 (0.003%)	—	—
Incinerator	177.38 (99.42%)	2866.52 (85.55%)	321.38 (61.75%)	5494.90 (99.27%)
Total	178.42 (100%)	3350.71 (100%)	520.17 (100%)	5536.63 (100%)

Throughput allocation of collection stations is shown in Table 4. Dongshih, Waipu, Da-an, and Heping collection stations have very few or no throughputs assigned, indicating that they may be closed to improve cost efficiency. Although composting is recognized as one of the safest and least harmful treatment methods for kitchen wastes, due to the fact that the net profit of compost is much lower than that of swine feed, all of the kitchen waste at each collection station is sold to feed swine. The throughputs, costs, revenue, amount of fly/bottom ashes, and pollution emissions generated by the three incinerators are shown in Table 5. The Houli and Wurih incinerators, whose WETCs are high, are operated at full capacity to achieve a larger electricity sale income, whereas throughput at the Nantun incinerator is only 406 ton/day (about 45% capacity) due to its low WETC. Table 3 also summarizes the amounts of SO_x, NO_x, CO, and PM emissions for Scenario 1. It was found that incinerators contribute the dominant pollutant emissions, especially SO_x and PM. Transportation is another major emission source of CO. Similar to conventional MSWM optimization problems, the major concern of Scenario 1 discussed here is the least-cost strategy. As the concerns of environmental impacts need to be included, the strategy is expected to be revised and the least-cost objective can no longer be maintained. Such a trade-off situation will be discussed in the scenario of pollution reduction in the following section.

Table 4. The throughputs among collection stations for Scenario 1

	Throughput (ton/day)
Collection Station	General Solid Waste	Resource Recyclable	Kitchen Waste	Total
Taichung	300	0	0	300 *
Taiping	145.82	39.05	15.13	200 *
Dali	144.40	42.50	13.10	200* *
Wufong	177.55	17.07	5.38	200 *
Wurih	200	0	0	200 *
Fengyuan	152.56	35.29	12.15	200 *
Houli	152.18	11.41	2.30	165.89
Shihgang	68.13	15.59	8.29	92.01
Dongshih	0	0	0	0^#
Heping	9.65	1.50	0.33	11.48
Sinshe	0	47.01	0	47.01
Tanzih	172.47	19.30	8.23	200 *
Daya	0	154.08	45.92	200 *
Shengang	184.14	12.65	3.21	200 *
Dadu	83.21	92.77	24.02	200 *
Shalu	90.89	81.95	27.16	200 *
Longjing	77.42	31.20	7.46	116.08
Wuci	50.21	33.99	3.17	87.37
Cingshuei	78.18	14.21	3.29	95.68
Dajia	72.83	15.19	5.30	93.32
Waipu	0	0	0	0^#
Da-an	0	2.13	0.37	2.50
Notes: *Full loaded; ^#No throughputs.

Table 5. The throughputs and pollution emissions among incinerators for Scenario 1

Incinerator	Nantun	Houli	Wurih	Total
Throughput (ton/day)	406	900	900	2206
Treatment cost (NT$/day)	198,694	305,100	116,100	619,894
Income (NT$/day)	213,996	710,528	754,287	1,678,810
Net cost (NT$/day)	−15,302	−405,428	−638,187	−1,058,917
Fly ash (ton/day)	20	45	54	119
Bottom ash (ton/day)	57	153	135	345
SOx (kg/day)	24	108	45	177
NOx (kg/day)	585	1305	1107	2997
CO (kg/day)	24	63	234	321
PM (kg/day)	122	1827	3546	5495

Results of Scenario 2

In Scenario 1, whose objective is only focused on minimizing the net cost, the incinerators with lower net operation costs such as Houli and Wurih are prioritized to treat the MSW, although they emit more air pollutants. However, when the pollutant emissions require further reduction, the throughputs of MSW will be shifted to the Nantun incinerator, which has lower emission but higher net operation costs compared with Houli and Wurih. Due to the model feasibility, the maximum emission reduction rates, based on the results of Scenario 1, of SO_x, NO_x, CO, and PM are 16%, 6%, 20%, and 32%, respectively. No feasible solutions can be found if further reduction rates are required in the optimization model. Table 6 summarized the comparisons of costs for the least-cost strategies with different emission reduction requirements in Scenario 2. It was found that the net costs increase as the emission reduction rates increase.

Table 6. Variation of related costs and benefits for different reduction levels of each pollution emissions

Cost/Benefit	Regulated Pollutant	Scenario 1	Reduction Rate of Pollution Emission
			5% (2%)*	10% (4%)*	15% (6%)*	Max (%)
Net cost (NT$/day)	SOx	−712,510	−668,715 (+6%)	−620,879 (+13%)	−568,025 (+20%)	−557,274 (+22%)
	NOx		−671,480 (+6%)	−622,606 (+13%)	−563,968 (+21%)	−563,968 (+21%)
	CO		−651,100 (+7%)	−587,213 (+18%)	−513,967 (+29%)	−431,097 (+39%)
	PM		−670,800 (+6%)	−628,899 (+12%)	−585,768 (+18%)	−426,000 (+40%)
Transportation cost (NT$/day)	SOx	357,528	350,085 (−2%)	346,607 (−3%)	348,094 (−2.6%)	348,570 (−2.5%)
	NOx		350,674 (−1.9%)	346,626 (−3%)	348,273 (−2.6%)	348,273 (−2.6%)
	CO		348,763 (−2%)	341,893 (−4%)	339,389 (−5%)	339,927 (−4.9)
	PM		351,730 (−2%)	346,120 (−3%)	341,713 (−4%)	338,877 (−5%)
Treatment cost (NT$/day)	SOx	3,031,469	3,053,726 (+1%)	3,076,015 (+1%)	3,098,328 (+2%)	3,102,792 (+2%)
	NOx		3,052,268 (+1%)	3,075,257 (+1%)	3,100,013 (+2%)	3,100,013 (+2%)
	CO		3,072,282 (+1%)	3,113,230 (+3%)	3,155,991 (+4%)	3,203,305 (+6%)
	PM		3,058,770 (+1%)	3,086,074 (+2%)	3,113,393 (+3%)	3,206,131 (6%)
Disposal cost (NT$/day)	SOx	1,014,455	1,004,439 (−1%)	994,409 (−2%)	984,368 (−3%)	982,359 (−3%)
	NOx		1,005,095 (−1%)	994,750 (−2%)	983,610 (−3%)	983,610 (−3%)
	CO		1,010,248 (0%)	1,005,656 (−1%)	998,890 (−2%)	992,975 (−2%)
	PM		1,011,042 (−0.3%)	1,007,629 (−0.6%)	1,004,214 (−1%)	991,912 (−2%)
Revenue (NT$/day)	SOx	5,115,961	5,076,964 (−1%)	5,037,910 (−2%)	4,998,815 (−2%)	4,990,994 (−2%)
	NOx		5,079,517 (−1%)	5,039,239 (−1%)	4,995,864 (−2%)	4,995,864 (−2%)
	CO		5,082,392 (−1%)	5,047,991 (−1%)	5,008,236 (−2%)	4,967,305 (−3%)
	PM		5,092,342 (−0.5%)	5,068,722 (−1%)	5,045,088 (−1%)	4,962,921 (−3%)
Notes: *Percentages in parentheses represent the reduction rates of NOx.

The MSW throughputs of the three incinerators illustrated in Figure 3 show that MSW originally incinerated by Houli needs to be shifted to Nantun to reduce the SO_x or NO_x emissions. On the other hand, Wurih incinerator's MSW must be shifted to Nantun if reduction of CO or PM emissions is required. As shown in Figure 4, such MSW shifts also result in the throughputs of ash disposal landfills near the Houli incinerator (namely the Houli and Shegang landfills) and Wurih incinerator (namely the Dali landfill) being decreased and these of the landfills near the Nantum incinerator, namely Taichung landfill and Wufong landfill, being increased. Furthermore, since some MSW are transferred to the nearer collection stations and incinerators to reduce the emissions of transportation, the transportation cost is slightly decreased.

Figure 3. Variation of throughputs of incinerators for different reduction levels of pollutant emissions.

Figure 4. Variation of throughputs of landfills for different reduction levels of pollutant emissions.

Table 7 summarizes the emission variations of the air pollutants while the emission reduction of a specific pollutant is required in the optimization model. It was found that as SO_x emission was reduced, PM emission was reduced about the same percentage, whereas the reductions of NO_x and CO were not so significant. When NO_x emission is reduced by r%, approximately more than a 2r% reduction of the emissions of SO_x and PM will also be achieved. However, the reduction of CO and PM both result in a slight increase of SO_x emission, and the reduction percentage of PM keeps steadily at 1.5 times CO's emission reduction rate. Furthermore, the unit cost of emission reduction of SO_x is the highest, whereas PM is the lowest. Based on the aforementioned analysis, PM can be used as the emission reduction indicator of the MSWM due to the fact that its reduction can also results in remarkable emission reductions of other pollutants with the exception of SO_x.

Table 7. Variation of pollution emissions for reducing a specific air pollutant

	Amount of Pollution Emission (kg/day)
Regulated pollutant		Original Emission (Scenario 1)	5% Reduction	10% Reduction	15% Reduction	Max Reduction (16%)
SOx	SOx	178	169 (−9*, −5%^#)	161 (−17, −10%)	152 (−26, −15%)	150 (−28, −16%)
	NOx	3351	3280 (−71, −2.1%)	3215 (−136, −4%)	3154 (−197, −5.8%)	3152 (−199, −5.9%)
	CO	520	516 (−4, −0.7%)	513 (−7, −1.3%)	512 (−8, −1.5%)	512 (−8, −1.5%)
	PM	5536	5278 (−258, −4.7%)	5021 (−515, −9.3%)	4764 (−772, −13.9%)	4712 (−824, −14.9%)
	Unit reduction cost (NT$/ton)		4910	5136	5399	5438
		Original Emission (Scenario 1)	2% Reduction	4% Reduction	6% Reduction	Max Reduction (6%)
NOx	SOx	178	170 (−8, −4.5%)	161 (−17, −9.6%)	151 (−27, −15.2%)	151 (−27, −15.2%)
	NOx	3351	3284 (−67, −2.0%)	3217 (−134, −4.0%)	3150 (−201, −6.0%)	3150 (−201, −6.0%)
	CO	520	516 (−4, −0.77%)	513 (−7, −1.3%)	512 (−8, −1.5%)	512 (−8, −1.5%)
	PM	5536	5295 (−241, −4.4%)	5030 (−506, −9.1%)	4744 (−792, −14.3%)	4744 (−792, −14.3%)
	Unit reduction cost (NT$/ton)		612	671	739	739
		Original Emission (Scenario 1)	5% Reduction	10% Reduction	15% Reduction	Max Reduction (20%)
CO	SOx	178	180 (+2, +1.1%)	181 (+3, +1.7%)	182 (+4, +2.2%)	183 (+5, +2.8%)
	NOx	3351	3314 (−37, −1.1%)	3279 (−72, −2.1%)	3243 (−108, −3.2%)	3213 (−138, −4.1%)
	CO	520	494 (−26, −5%)	468 (−52, −10%)	442 (−78, −15%)	416 (−104, −20%)
	PM	5536	5131 (−405, −7.3%)	4722 (−814, −14.7%)	4274 (−1236, −22.3%)	3796 (−1740, −31.4%)
	Unit reduction cost (NT$/ton)		2361	2409	2545	2705
	Original Emission (Scenario 1)		5% Reduction	10% Reduction	15% Reduction	Max Reduction (32%)
PM	SOx	178	179 (+1, +0.56%)	180 (+2, +1.1%)	181 (+3, +1.7%)	182 (+4, +2.2%)
	NOx	3351	3324 (−27, −0.8%)	3298 (−53, −1.6%)	3274 (−77, −2.3%)	3204 (−147, −4.4%)
	CO	520	503 (−17, −3.3%)	485 (−35, −6.7%)	468 (−52, −10%)	416 (−104, −20%)
	PM	5536	5259 (−277, −5%)	4982 (−554, −10%)	4705 (−831, −15%)	3764 (−1772, −32%)
	Unit reduction cost (NT$/ton)		151	151	153	162
Notes: *Change amount of pollution emission relative to original emission. ^#Change percentage of pollution emission relative to original emission.

Conclusion

A prototype LP optimization model was developed to analyze the least-cost strategies of MSWM in Taichung City. Air pollutant emissions during the life cycle of MSWM were also evaluated. The results show that currently the MSWM of Taichung City can achieve a negative net cost. Resource recycling is the most economically efficient MSW treatment method that generates most of the net profits. However, there are no significant benefits for MSW treated in incinerators. All of the kitchen waste is sold as swine feed, and currently composting is not an economically attractive treatment method for kitchen waste. The net cost of MSWM significantly increases if the air pollutant emission reduction is implemented. The available emission reduction percentages of SO_x, NO_x, CO, and PM are 16%, 6%, 20%, and 32%, respectively. PM can be used as the emission reduction indicator for the MSWM system, since its reduction can also result in remarkable emission reduction of other pollutants. The model developed in this work may aid decision makers to set up least-cost and more environment friendly MSWM strategies.

The advantage of this optimization model is the ability to calculate the least-cost strategy and air pollutant emission of the life cycle at the same time. The decision maker can further explore variation of cost and profit of optimization strategy while different emission reduction of the air pollutants is implemented in the MSWM system. The current consideration of the pollution emission for this model is restricted to four air pollutants due to the fact that real-world data of other pollutants are insufficient at the moment. In addition, the optimization model of this study is a LP-based model with a single objective function. The multiobjective programming techniques can be used to obtain a trade-off solution between economical factors and environmental impacts in a future study.

Acknowledgments

This research was funded by a grant from the National Science Council of the Republic of China (NSC-99-2221-E-005-035-MY3). The authors gratefully appreciate this support.

Appendix: List of Notations

Table

Definition of decision variables
BASH jl	the amount of bottom ash transferred from incinerator j to landfill l
BASH je	the amount of bottom ash transferred from incinerator j to reuse facilities e
BASHGR j	the generation rates of bottom ash in incinerators
EM z	the total emissions of pollutant z (z = CO, NOx, SOx, or PM) for MSWM system
FASH jl	the amount of fly ash transferred from incinerator j to landfill l
FASHGR j	the generation rates of fly ash in incinerators j
GSW ik	the amount of GSW collected from district i and sent to collection station k
KW ik	the amount of kitchen waste collected from district i and sent to collection station k
KW kf	the amount of kitchen waste transferred from collection station k to compost facility f
MIXSW kj	the amount of GSW mixed with the rejected materials sorted from recyclable materials from collection station k and transferred to incinerator j
REVE	the total revenue
REC ik	the amount of recyclable materials collected from district i and sent to collection station k
RESFW fj	the amount of residual materials from compost facility f transferred to incinerator j
RESBASHR el	the amount of the residual materials transferred from bottom ash reuse facility e to landfill l
RESKWGR f	the generation rate of residuals for compost
RESBASHRR e	the generation rate of residuals for bottom ash reused
TRAN_COST	the total transportation cost including the collection and transferring cost
TREAT_COST	the total treatment cost
Definition of input parameters
α i	the ratios of recyclable materials to MSW in district i
β i	the ratios of kitchen waste to MSW in district i
BASHC l	the unit treatment costs for bottom ash of landfill l
BASHC e	the unit treatment cost for bottom ash of reuse facility e
CC	the unit collection cost
COMPGR f	the compost generation rate of compost facility f
COMPSP f	the selling price of compost of compost facility f
CAP k	the designed capacities for collection station k
CAP j	the designed capacities for incinerator j
CAP f	the designed capacities for compost facility f
CAP l	the designed capacities for landfill l
CAP e	the designed capacities for bottom ash reuse facility e
CAR C	the capacities of collection vehicles
CAR T	the capacities of transportation vehicles
CAR C	the capacities of collection vehicles
CAR T	the capacities of transportation vehicles
D xy	the distance from x (x = i, k, j, f, e) to y (y = k, j, f, l, e)
EP j	the selling price of electricity
EMC z	the emission factors of pollutant z for collecting MSW
EMT z	the emission factors of pollutant z for transporting MSW
EMF_INC jz	the emission factors of pollutant z for incineration
EMF_KW fz	the emission factors of pollutant z for composting
EMF_SFD kz	the emission factors of pollutant z for feeding swine
FASHC l	the unit treatment costs for fly ash of landfill l
GENE i	the MSW generated in district i
KWCC f	the unit treatment cost for kitchen waste at compost facility f
RR ir	the ratio of category r material to all recyclable materials of district i
RRC r	the unit treatment cost of material r
RRSP r	the selling price of recyclable material r
SFD k	the amount of kitchen waste for feeding swine of collection station k
SFDC k	the unit treatment cost of kitchen waste for feeding swine of collection station k
SFDSP k	the selling price of kitchen waste for swine feeding
SR j	the proportion of electricity generated by incinerator j which is sold
TC	the unit transferring cost
TC j	the unit treatment cost of incinerator j
WETC j	the waste-to-electricity transfer coefficient (WETC)