Waste Management 43 (2015) 9–18

Contents lists available at ScienceDirect

Waste Management
journal homepage: www.elsevier.com/locate/wasman

Optimization of municipal solid waste collection and transportation

routes
Swapan Das ⇑, Bidyut Kr. Bhattacharyya
Department of Mechanical Engineering, Indian Institute of Engineering Science and Technology, Shibpur, Howrah 711103, India

a r t i c l e i n f o a b s t r a c t

Article history: Optimization of municipal solid waste (MSW) collection and transportation through source separation
Received 31 March 2015 becomes one of the major concerns in the MSW management system design, due to the fact that the
Revised 8 June 2015 existing MSW management systems suffer by the high collection and transportation cost. Generally, in
Accepted 17 June 2015
a city different waste sources scatter throughout the city in heterogeneous way that increase waste col-
Available online 4 July 2015
lection and transportation cost in the waste management system. Therefore, a shortest waste collection
and transportation strategy can effectively reduce waste collection and transportation cost. In this paper,
Keywords:
we propose an optimal MSW collection and transportation scheme that focus on the problem of minimiz-
Solid waste
Collection
ing the length of each waste collection and transportation route. We first formulize the MSW collection
Transfer and transportation problem into a mixed integer program. Moreover, we propose a heuristic solution for
Transportation the waste collection and transportation problem that can provide an optimal way for waste collection
Route optimization and transportation. Extensive simulations and real testbed results show that the proposed solution can
significantly improve the MSW performance. Results show that the proposed scheme is able to reduce
more than 30% of the total waste collection path length.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction problems to develop an integrated waste management system

(Nuortio et al., 2006). In a bid to reduce collection costs as well
The rapid and constant growth in urban population led to a dra- as transportation cost, municipal authorities have been motivated
matic increase in municipal solid waste (MSW) generation, with a to develop new strategies for the collection and transportation of
crucial socioeconomic and environmental impact. Municipal solid solid wastes, particularly in urban centres (Khan and Samadder,
waste management is a multidisciplinary activity that includes 2014; Kanchanabhan et al., 2011; Rada et al., 2010).
generation, source separation, storage, collection, transfer and Generally, collection and transportation are the most important
transportation, processing and recovery, and, last but not least, dis- and costly aspect of the process because of the labour intensity of
posal (Rada et al., 2013; Bovea et al., 2010; Gallardo et al., 2015; the work and the massive use of vehicles in the collection and
Rada, 2014; Minoglou and Komilis, 2013; Wagner and Bilitewski, transportation process (Amponsah and Salhi, 2004). In India, waste
2009). In order to develop a sound material-cycle society, cost- collectors start the journey from a collection centre for collection of
effective integrated municipal solid waste management system is segregated waste and visit the various sources with containerized
required for the municipalities (Weng and Fujiwara, 2011; tri-cycle (hand cart). The containerized tri-cycle having six buckets
Consonni et al., 2011; Massarutto et al., 2011; Ionescu et al., of 40–50 L capacity. After the full fill the capacity of tri-cycle bar,
2013; Eriksson et al., 2014). Solid waste collection is widely recog- they conveyed the garbage to the collection centre. Similarly, a
nized across the globe to account for a majority of expenditure on vehicle starts from the depot, visits a number of collection cen-
solid waste management (Jacobsen et al., 2013). Typically, collec- tre/transfer station and ends at the depot (Kim et al., 2006).
tion costs represent 80–90% and 50–80% of municipal solid waste Depending on the complexity of the problem and different waste
management budget in low income and middle income countries characteristics, different types of vehicles are used for collection
respectively (Aremu, 2013). Therefore, waste collection and trans- and transportation (Wy et al., 2013). An integrated waste manage-
portation problems are one of the most difficult operational ment system can improve the vehicle routing, dispatching, mainte-
nance and management (Sahoo et al., 2005; Xu et al., 2015; Li et al.,
⇑ Corresponding author. 2014; Ionescu and Stefani, 2014; Rabbani et al., 2014) and in this
E-mail addresses: swapan2009sajal@gmail.com (S. Das), bidyut53@yahoo.co.in paper, in order to improve the road compactness and workload
(B.Kr. Bhattacharyya). balancing, a scheme is developed for municipal solid waste

http://dx.doi.org/10.1016/j.wasman.2015.06.033
0956-053X/Ó 2015 Elsevier Ltd. All rights reserved.
10 S. Das, B.Kr. Bhattacharyya / Waste Management 43 (2015) 9–18

collection and transportation: an optimization scheme is needed to the segregated solid waste from the transfer station to nearby
reduce waste collection and transportation route length. categorized waste processing plant; (iv) transfer the produced
This paper mainly focuses on the collection, transfer and trans- waste from processing plant to the nearest landfill site. Each part
portation of solid waste from any waste generation sources contents huge amount of cost for waste collection or waste mate-
(households, markets, institutions and offices) to the processing rial transportation. The proposed scheme optimizes transportation
plant or landfill site. The factor influencing the municipal solid cost of each part of the system. Therefore, the total waste manage-
waste collection and transportation are: (i) quantity of municipal ment cost for the whole system is optimized. The optimization pro-
solid waste generation; (ii) number of waste collectors and their cess for each waste management part is described in Fig. 1.
deployment to cover the whole city; (iii) vehicles to run to and
from various transfer stations where garbage is gathered to the
3.1. Optimal waste collection from source to collection centre
respective processing plant. In this paper a scheme is proposed
for optimizing municipal solid waste collection and transportation
In phase 1, waste collectors collect waste from different sources
routes using the Travelling Salesman Problem (TSP). The proposed
in different areas and transfer to a collection centre (shown in
scheme design an optimized municipal solid waste management
Fig. 2). Here, we find optimize route between each collection centre
(MSWM) system with respect to transportation route length. The
and its sources. The optimal route between the collection centre
TSP is one of the most prominent problems in combinatorial opti-
and its sources, reduces waste collection cost in phases one. We
mization, and at the same time a quintessential applied spatial-
represent waste collection point through a connected graph
analytic challenge. The well-known travelling salesman problem
G (V, E), where V = {v1, v2, . . ., vn} (n – 0) is the number of sources
is the following: a salesman is required to visit once and only once
point and E = {e1, e2, . . ., en} (n – 0) is the number of edges that con-
each of n number of different cities starting from a base city, and
nect all source points.
returning to this city. The task is to find a shortest possible tour
In phase 1 optimization, m number of waste collectors visit each
that visits each city exactly once by the salesman (Dumitrescu
vi centre where vi e V through optimal path sequence, where ei e E.
et al., 2009; Renaud et al., 2002).
Our optimal waste collection path calculation problem is related to
travelling salesman problem (TSP). Therefore, our objective func-
2. Problem formulation tion for optimal waste collection path calculation is

In waste management system, n number of sources are avail- X

n X
n

able in different municipalities under a city. Each source point gen-

Minimize v ij xij ð3Þ
i¼1 j¼1;j–i
erates some waste material in each day. The waste management
Subject to
system collects these waste form the different sources and process
it through different plants. Therefore, total waste management cost Xn
xij ¼ m ð4Þ
(Wmanagement(Cost)) for n number of sources is j¼2
X
n
X
n X
n X
n X
k
xi1 ¼ m ð5Þ
W management ðCostÞ ¼ Cc þ Ct þ Cp þ Cd þ C ð1Þ i¼2
i¼1 i¼1 i¼1 i¼1
Xn
xij ¼ 1; 8j ¼ 2;. ..;n ð6Þ
where Cc is the collection cost, Ct is the transportation cost, Cp is the i¼1;i–j
processing cost, Cd is the disposal cost for k number of unused or X
n

produced waste material after processing and C is the constant cost xij ¼ 1; 8i ¼ 2;...;n ð7Þ
that are depending on other parameter such as accident, mainte- j¼1;i–j

nance of collection centre, transfer station and vehicle. The profit xij 2 f0;1g; 8i;j ¼ 1; ...;n;i–j ð8Þ

gain by the n number of sources is 1; if source i directly procedes source j on any tour
xij ¼
X
f 0; otherwise; 8i;j ¼ 1;.. .;n; i–j;
W management ðprofitÞ ¼ PðRevenueÞ  W management ðCostÞ ð2Þ
Our main objective is to minimize waste collection path cost in
i¼1
the lower level where each waste collector collects waste from dif-
where P(Revenue) is mainly achieved from the sales of recyclable ferent sources. Constraints (4) and (5) insure that m number of
materials, sales of compost product and the sales of electricity. A waste collector starting from a source points and visits each source
waste management system always tries to develop a mechanism point and return to starting point. Constraint (6) and (7) are the
that increases the profit (Wmanagement(profit)). Therefore, the objec- classical assessment restriction that each source point in the set
tive of any waste management scheme is how to reduce waste col- {v1, v2,. . ., vn} to be visited by at least one waste collector and at
lection cost and waste processing cost. In this paper, we increase one connection has been existed between the neighbour source
P P
Wmanagement(profit) to reduce ni¼1 C c and ni¼1 C t . point. The logical representation of the decision variables are
Therefore, our main objective is to identify optimal waste col- represented in (8).
lection and transportation routes for reduction of waste collection
and transportation cost in the waste management system design. 3.2. Optimal waste transfer from collection centre to transfer station

3. Proposed scheme In phase 2, collected waste is transferred from the collection

centre to transfer station through different small vehicle (shown
The proposed scheme divides whole integrated waste manage- in Fig. 3). The optimized route between collection centre and trans-
ment system into four different parts (shown in Fig. 1). (i) Collec- fer station minimize waste transportation cost. Transfer stations
tion of segregated solid waste from various sources (households, and collection centres are represented by an undirected connected
markets, offices and institutes) and conveyed waste to the nearest graph T(C, D) where, C = {c1, c2, . . ., cn} (n – 0) are all points that
collection centre; (ii) transfer accumulate solid waste from vehicles are visited for waste transportation and D = {d1, d2, . . .,
collection centre to the adjoining transfer station; (iii) transport dn} (n – 0) are the double way route between them.
 S. Das, B.Kr. Bhattacharyya / Waste Management 43 (2015) 9–18 11

Step 1
Sources

Household Market Office Institute

Step 2
Collection Centre-1 Collection Centre-2 Collection Centre-3

Step 3 Transfer Station-1 Transfer Station-2

Recycling Composting Waste to Energy

Step 4

Landfill Site-1 Landfill Site-2

Fig. 1. Architectural view for MSW collection and transportation.

Fig. 2. Different road between sources and collection centres.

In phase 2, we calculate the optimal path between the collec- optimization problem we consider the number of visits in each col-
tion centre and transfer station. In the second level optimization, lection centre must be once. In second waste collection tour, each
g number of the transfer vehicles, visit each ci collection centre collection centre may be visited k number of times, if the amount
where ci e C through the optimal path sequence and di e D. In this of waste is more than the capacity of a transfer vehicle. We relate
12 S. Das, B.Kr. Bhattacharyya / Waste Management 43 (2015) 9–18

Fig. 3. Road between collection centres (red mark) to transfer stations. (For interpretation of the references to colour in this figure legend, the reader is referred to the web
version of this article.)

WtE & Recycling Plant

Transfer Station
Composting
Plant

Transfer Station Composting

Plant
Transfer Station
WtE Plant
Composting
Transfer Station Plant

Composting Plant Recycling Plant

WtE & Recycling Plant

Fig. 4. The optimal transportation of waste material from transfer stations to processing plants.

(
our optimal path problem with covering salesman problem (CSP). 1; if collection centre i is on the tour;
The objective function for optimal path calculation is computed wi ¼
0; otherwise
as follows:
(
 1; if arc ði; jÞ is chosen to be in the solution;
1; if collection centre j can cover collection centre i; xij ¼
aij ¼ 0; otherwise
0; otherwise
We introduce the decision variables: The integer programming model can be now stated as
 S. Das, B.Kr. Bhattacharyya / Waste Management 43 (2015) 9–18 13

X X
Minimize cij xij þ F i wi ð9Þ where L is the connecting route between transfer stations and pro-
ði;jÞ2A i2N cessing plants. The objective is to minimize the sum of the tour costs
Subject to and transfer visiting costs. Constraint set (16) ensures that if transfer
X X station i is visited yi times, then we have yi incoming and yi outgoing
xji ¼ xij ¼ wi 8i 2 N ð10Þ
arcs. Constraint set (17) specifies the demand of each transfer station
j:ðj;iÞ2A j:ði;jÞ2A
X must be covered. Constraints set (18) and (19) are linking con-
aij wi  ki 8i 2 N ð11Þ straints, ensuring that wi and xij are 1 if yi or zij are greater than 0
j2N
XX XX (i.e. if a transfer station is visited or an arc is traversed). Note that
xlk þ xkl  2ðwi þ wj  1ÞS  N; it suffices to set L = maxieN{ki}. Constraint set (20) is a connectivity
l2S k2NnS k2NnS l2S constraint that ensures that there are no sub tours.
2  jSj  n  2; i 2 S; j 2 N n S ð12Þ Note again, that there are an exponential number of connectiv-
xij 2 f0; 1g 8i; j 2 A ð13Þ ity constraints. Finally, constraints set (21) and (22) define the vari-
ables as binary an integer as appropriate. For the ICSP with
wi 2 f0; 1g 8i 2 N: ð14Þ
overniths, the above integer programming model is valid if we aug-
The objective is to minimize the sum of the tour costs and col- ment the arc set A with a self-loops. Specially, we add to A the arc
lection centre visiting costs. Constraint set (10) ensures that for set {(i, i): i e N} or (A = A [ {(i, i): i e N}) with cii the cost of self-loop
each on-tour collection centre, we have one incoming and one out- arcs (i, i) set to 0.
going route. Constraint set (11) specifies that the demand of each
collection centre must be covered. Constraint set (12) is a connec- 3.4. Optimal waste transportation from processing plant to landfill site
tivity constraints that ensures there are no sub tours. Note that
there are an exponential number of connectivity constraints. Con- In this phase we have not used any standard optimization tech-
straints (13) and (14) define the variable as binary. nique for transporting of producing waste from processing plant to
landfill sites because in real life scenarios the produced waste
3.3. Optimal waste transportation form transfer station to processing material is used as road construction material by the third party
plant or sent to landfills. Therefore, different third party contractors
carry this material from the plants. Therefore, we only set our pro-
In a practical scenario, processing plants is setup at outside of the cessing plant in the nearby a landfill site for despatching produced
city (shown in Fig. 4). Therefore, we use heavy vehicle to transfer (inert) waste usually. Therefore, unused materials are disposed
segregated waste to their respective processing plants. The optimal without any cost.
road length minimizes the transportation cost between the transfer
stations to processing plants. Transfer stations and processing plants
are represented by P(K, L), where, K = {k1, k2, . . ., kn} (n – 0) and 4. Methodology
L = {l1, l2, . . ., ln} (n – 0) are represented transfer stations and pro-
cessing plants and connection route between them. In waste management system, waste collection, transfer and
In the third stage, we minimize path length between the trans- transportation are a complex problem where three different pro-
fer station and processing plant. At third level, optimization h cesses are going on sequentially in three different levels. Each level
number of transport vehicle visit each ki transfer station where ki - uses individual collection or transportation strategy to collect and
e K through optimal path sequence, where li e L. Our optimal waste transport wastes from sources to processing plants. For this com-
transportation path calculation problem is related to Integer plex problem a local optimization problem is more cost effective.
Covering Salesman Problem without Overnights (ICSP). Therefore, we have used local optimization technique for each
In this stage, any vehicle may visit more than one time to a individual part that give better results for each particular real life
transfer station, because of in high population density city’s trans- problem.
fer station may contain the more amount of waste. Therefore, our Fig. 5 shows a schematic representation of the phases of the
objective function is methodology proposed. The first phase of the methodology consists
of number of sources, which are scattered throughout the city by
yi = number of times that transfer station i is visited by the tour, and heterogeneous way. Each of the source is connected to its neighbour
zij = number of times road (i, j) is traversed by the tour. sources by at least one route or more than one route. In the first
phase, we have pointed out all sources and all possible connecting
The integer programming model can now be stated as: routes between them according to existing map. Then we apply
X X our proposed route optimization technique to find out the optimal
Minimize cij zij þ F i yi ð15Þ routes for waste collection. In the second phase, our main objective
ði;jÞ2A i2N
is to transport waste from collection centres to transfer stations
Subject to through optimal routes. Therefore, the location of the collection cen-
X X
zji ¼ zij ¼ yi 8i 2 N ð16Þ tres and routes have been selected according to the existing map.
j:ðj;iÞ2A j:ði;jÞ2A Then we apply our proposed scheme for second phase optimization.
X In the third phase, the location of the transfer stations and routes
aij yj  ki 8i 2 N ð17Þ
between them has been identified through the detail existing map
j2N
study. Then we found out the optimal route between the transfer
yi  Lwi 8i 2 N ð18Þ stations by applying our third phase optimization process. The com-
zij  Lxij 8ði; jÞ 2 A ð19Þ bination of these three phase output shows the optimal way of
XX XX
xlk þ xkl  2ðwi þ wj  1ÞS  N; waste collection, transfer and transportation.
l2S k2NnS k2NnS l2S

2  jSj  n  2; i 2 S; j 2 N n S ð20Þ 5. Performance evaluation

xij 2 f0; 1g; zij 2 Z þ 8ði; jÞ 2 A ð21Þ
This section shows the simulation result with respect to real life
wi 2 f0; 1g; yi 2 Z þ 8i 2 N ð22Þ
data that have been gathered in the Kolkata metropolitan city (In-
14 S. Das, B.Kr. Bhattacharyya / Waste Management 43 (2015) 9–18

Number of source Number of Number of transfer

collection centre station

All possible route All possible route All possible route

between the sources between the between the
collection centres transfer stations
Optimum way of
Calculate optimal waste collection,
Calculate optimal Calculate optimal
route between the route between the transfer and
route between the
sources transfer stations transportation
collection centres

Collection of waste Transfer of waste Transportation of

by the optimal route by the optimal route waste by the
optimal route

Phase-1 Phase-2 Phase-3

Fig. 5. Schematic representation of the phases of the methodology proposed.

dia). The factors related to this proposed scheme are the Kolkata through GPS like devices. For simulation, we have imported the
metropolitan city area comprising about 1886.67 km2, having pop- real life data to the MATLAB simulation software. We calculated
ulation of 1,41,12,536 as per the 2011 census, density of 7480 per the optimal path between the 100 sources points through MATLAB
square km, quantity of waste generation 4837 tonne/day, per cap- with the help of the proposed scheme.
ita waste generation 0.35 kg/capita/day. Also considered the col- Table 1 shows the distance of each path between the sources.
lection centres capacity 2 tonne, transfer vehicle capacity According to these distances we calculated the optimal waste col-
1 tonne, transfer stations capacity 15–18 tonne and the transporta- lection path between the sources and collection centres. In real life,
tion vehicle capacity 4–5 tonne. In the performance evaluation Table 1 distance originally shows the each road length that has
process, the optimal path shows between the sources and waste been present between the sources.
collection centre, second we show the optimal path between the Fig. 6(b) shows the optimal path between the 100 source points.
waste collection centre and transfer station. After that, we describe The optimal path has been simulated in the MATLAB software
the optimal path between the transfer station to processing plant according to the real life data which are collected from ‘BF’ block
also shown in the figure. of the Bidhan Nagar municipal corporation area. The optimal path
reduces the tour length for waste collection. The path has been
5.1. Optimal waste collection between the sources and waste collection shown by the sequence of the sources that will be followed by
centre the waste collector. The sequence of the sources is following: 99
? 100 ? 95 ? 94 ? 93 ? 92 ? 89 ? 90 ? 89 ? 88 ? 87 ?
Fig. 6(a) shows all possible paths between the source points. 86 ? 85 ? 84 ? 83 ? 82 ? 81 ? 80 ? 57 ? 80 ? 79 ? 78
The position of the sources and path between the sources has been ? 77 ? 76 ? 75 ? 74 ? 91 ? 74 ? 73 ? 96 ? 97 ? 98 ?
identified according to ‘BF’ block of the Bidhan nagar municipal 71 ? 72 ? 66 ? 65 ?64 ? 63 ? 62 ? 61 ? 60 ? 59 ? 58
corporation area (under Kolkata metropolitan city). ? 56 ? 55 ? 54 ? 53 ? 50 ? 49 ? 48 ? 47 ? 43 ? 42 ? 43
Initially, we have taken 100 number of source points for our ? 68 ? 69 ? 70 ? 39 ? 40 ? 41 ? 35 ? 36 ? 37 ? 38 ?
experiment, we have identified distances between these sources 12 ? 13 ? 14 ? 11 ? 10 ? 17 ? 16 ? 15 ?16 ? 17 ? 18
according to their position. The source position has been identified ? 19 ? 34 ? 46 ? 45 ? 44 ? 33 ? 32 ? 31 ? 30 ?51 ? 52

(b)

Fig. 6. (a) All possible path between the sources. (b) Optimal path between the sources.
 S. Das, B.Kr. Bhattacharyya / Waste Management 43 (2015) 9–18 15

Table 1
Path distance (m) between the sources.

Source Destination source Distance Source Destination source Distance Source Destination source Distance
1 2 7 31 32 7 63 64 12.37
1 26 16 32 33 21 63 75 13.6
2 3 8 32 48 22 64 65 13
2 24 12 33 35 11.18 65 66 6.4
3 4 9 33 44 9 65 73 14.32
4 5 8 34 46 14 66 67 5.66
5 6 7.21 35 36 13.04 66 72 9.85
6 7 6 35 41 13.93 67 68 15
6 21 10.2 36 37 12 67 69 10
7 8 11.4 36 40 11 69 70 10.63
8 9 5 37 38 8.94 69 71 8
9 10 7.21 37 39 12.53 71 72 5
9 17 12 39 40 6 71 98 12
9 19 14.56 39 70 7.81 73 74 19.24
9 20 11 40 41 8.06 73 96 11.18
10 11 31.62 41 42 7.81 73 98 25.02
10 17 7.21 42 43 8.49 74 75 7.21
11 12 16.12 43 44 13.6 74 91 10.77
11 14 15.23 43 47 15 75 76 4.24
12 13 8 43 63 6.4 76 77 4.12
12 38 14 43 64 11.31 77 78 7.28
13 14 6 43 68 15 78 79 7
14 15 7 44 45 4.12 79 80 8.94
15 16 4.47 45 46 4 79 87 12
15 36 18.03 46 47 12.37 80 81 8.06
16 17 7.28 47 48 10 81 82 9.9
17 18 9.22 47 60 9 82 83 8.49
18 19 14.87 47 62 7.62 82 84 17
18 35 8.06 48 49 5 83 84 12.53
19 33 6.4 49 50 15 84 85 6.08
19 34 4.12 50 51 10.05 85 86 9
20 21 10 50 53 14 86 87 8.25
21 22 12 50 55 7.21 87 88 8.6
21 32 19 51 52 8.25 88 89 7.28
22 23 7 52 82 26 89 90 4.47
23 24 8.94 53 54 5.39 89 92 14
23 25 8 54 55 8.06 90 91 5.83
24 25 4 54 82 13.93 92 93 8
25 26 7 55 56 11 93 94 7.28
25 29 9 56 57 9.06 94 95 7.28
26 27 22 56 58 10.2 95 96 10.05
27 28 9.49 57 80 11 95 100 10
27 53 19 58 59 9 96 97 13.89
28 29 10.2 59 60 4.12 97 98 8.94
28 30 8 60 61 4 97 99 18.79
28 52 14.14 60 77 11 99 100 9
30 31 10 61 62 6.71 – – –
30 51 12.17 62 63 4.47 – – –

Fig. 7. (a) All possible path between the collection centres. (b) Optimal path between the collection centres.
16 S. Das, B.Kr. Bhattacharyya / Waste Management 43 (2015) 9–18

Table 2
Path distance (km) between the collection centres.

Collection Destination collection Distance Collection Destination collection Distance Collection Destination collection Distance
centre centre centre centre centre centre
1 2 9 21 22 3.61 42 60 5.1
1 20 7 22 23 8.54 42 64 13.04
1 22 11.7 22 24 4.24 43 44 3.61
2 3 5.83 23 26 2.83 43 45 5.39
2 8 4 24 25 4.24 43 48 12.37
3 4 5.66 24 26 8.06 44 45 4.47
4 5 9.22 25 28 5.39 45 46 3
4 7 6.32 25 29 4.12 46 47 6.4
5 6 6.71 25 30 9.22 46 60 9.22
6 7 8.49 26 27 4.12 47 48 4.24
6 12 8.25 27 40 4.47 47 58 11.18
7 8 11.4 28 32 11.4 48 49 4.12
7 11 4 28 39 11 49 50 4.12
7 15 7.21 28 40 7.28 49 56 3
8 10 5 29 30 7.62 50 51 4.47
8 16 10 30 31 4.12 50 53 7.07
9 10 5 30 32 6 51 52 5
9 20 7.07 31 32 6.4 51 53 5.83
10 19 4 32 33 5.1 52 53 7.28
10 20 5 33 34 3.16 52 56 12.37
11 16 3.16 33 35 10.2 53 54 4.24
12 13 5.83 34 39 5 54 55 5.83
12 65 6.32 35 36 7.81 54 56 8.25
13 14 5 36 37 2.83 55 56 3.16
13 64 5.1 36 39 6 55 57 4
14 15 3.61 36 46 5.39 56 60 13.34
14 63 7.21 37 38 4.12 57 58 4.12
15 16 3.16 37 45 3.16 58 59 4.12
15 17 2.83 38 39 3 59 60 6
16 19 5 38 40 4.12 59 61 4.12
17 18 4.47 38 44 4.12 61 62 9.9
17 63 6.08 39 40 4.47 61 65 5.39
18 19 6.08 40 41 8.54 62 63 4.24
18 21 10.82 41 42 3.61 63 64 7.28
18 23 5.39 41 44 3.61 64 65 6.32
19 24 12.37 41 62 5.66 – – –
20 21 2.24 42 43 3.61 – – –

? 28 ? 27 ? 26 ? 25 ? 29 ? 24 ? 23 ? 22 ? 21 ? 20 ?9 ? transfer stations. This optimized path not only optimized waste

8 ? 7 ? 6 ? 5 ? 4 ? 3 ? 2 ? 1. The numbering of the sources collection cost it also minimizes waste collection time.
given as per the planning of Municipality Corporation (Bidhan The path cost between the collection centres are shown in
Nagar). The numbering of the sources are used to identify the Table 2. These costs indicate road distance between the collection
sequence of the optimal visiting path. The total path length in centres.
the case study area is 2947 m and the optimal path length is Fig. 7(b) show the optimal path between the 65 collection cen-
2009 m. The optimal waste collection path can reduce the number tres. This optimal path has been calculated according to the opti-
of waste collector if waste collection time is constant. On the other mization formula shows in Eq. (9). Small size vehicles, visit each
hand, if waste collectors are fixed, then waste collection time has waste collection centres through this optimal path and collects
been reduced. When waste collectors are fixed for a particular area, waste from the collection centres. After that collected waste has
then these waste collectors collect the waste from that particular brought to the transfer station. Initially, all vehicles start from an
area within the minimum time period. Therefore, we have used initial waste collection centre and each vehicle visit at least one
this collectors for loading and unloading purposes in collection waste collection centre before going to transfer station. On the
centres or transfer stations that can also reduce waste collection other hand optimal waste transfer, we also consider that all waste
labour cost. Therefore, the proposed scheme reduces marginally collection centres must be visited by at least one vehicle. The opti-
the overall collection cost of municipal solid waste. mization waste transfer path is represented by the sequence of the
collection centre. The sequence of the collection centres is follow-
5.2. Optimal path between waste collection centres and transfer ing: 52 ? 53 ? 51 ? 50 ? 49 ? 48 ? 49 ? 56 ? 54 ? 55 ? 57 ?
station 58 ? 47 ? 46 ? 60 ? 59 ? 61 ? 65 ? 12 ? 13 ? 64 ? 42 ? 43
? 44 ? 41 ? 44 ? 45 ? 37 ? 36 ? 35 ? 33 ? 34 ? 39 ? 38 ?
Fig. 7(a) shows the all possible paths between the 65 waste col- 40 ? 27 ? 26 ? 23 ? 18 ? 17 ? 63 ? 62 ? 63 ? 14 ? 15 ? 7 ?
lection centres. In our simulation, we have considered initially 65 6 ? 5 ? 4 ? 7 ? 11 ? 16 ? 19 ? 24 ? 25 ? 28 ? 32 ? 31 ? 30
waste collection centres available within the Bidhan Nagar munic- ? 25 ? 29 ? 25 ? 24 ? 22 ? 21 ? 20 ? 9 ? 10 ? 8 ? 2 ? 3 ?
ipal corporation area (under Kolkata metropolitan city). In real life, 2 ? 1. The total road length in the case study area is 19.89 km and
each path between the waste collection centres indicates route the optimal road length in the case study area is 12 km. The opti-
between the waste collection centres. According to our proposed mal path length can reduce the number of waste transfer vehicles
scheme, small size vehicles collects waste from the collection cen- (small size vehicles) and also reduces the waste transfer time.
tres and conveyed it to transfer station. In this process, we opti- Therefore, this optimization path marginally reduces the transfer
mize waste collection path between the collection centres and cost of municipal solid waste.
 S. Das, B.Kr. Bhattacharyya / Waste Management 43 (2015) 9–18 17

Fig. 8. (a) All possible path between the transfer stations. (b) Optimal path between the transfer stations.

Table 3
Path distance (km) between the transfer stations.

Transfer Destination transfer Distance Transfer Destination transfer Distance Transfer Destination transfer Distance
station station station station station station
1 2 7.28 16 39 12.37 31 32 4.12
1 4 9.49 17 18 3.61 32 33 4.47
2 3 3.61 17 24 5.1 33 34 5.39
2 4 16.03 18 19 3 33 35 5.39
3 5 10.2 18 20 4.47 34 43 9.06
3 7 5 19 20 4.12 35 36 8.06
3 8 5 19 23 5.1 35 43 6.32
3 9 10.05 20 21 5.83 36 42 2.83
4 5 5.66 20 22 3 37 39 9.06
5 6 4 21 22 5 37 41 6
6 7 6.71 21 27 4.47 38 39 5.39
6 18 3.61 21 33 11.4 38 42 7.07
7 10 5 22 23 4.24 38 44 4
8 9 6.32 22 24 6.32 39 40 6.4
8 10 5 22 27 5 40 44 9.06
9 11 8.54 23 25 3.61 40 47 11.4
10 11 13.34 23 31 9.9 40 50 11.18
10 18 12.37 24 25 3 41 44 4.12
11 12 4.12 26 27 6.32 41 46 4.47
12 13 9.06 26 29 5 42 43 12.37
13 14 7.07 26 30 5.39 43 46 13.34
13 16 8.25 27 28 3.16 44 45 3.16
14 15 4.12 28 29 5 45 46 5
15 16 6.08 28 36 8.25 45 47 3.16
15 17 3.16 29 30 5.1 46 48 3.16
16 24 5 29 31 4 47 49 2.83
16 25 5.83 29 37 3.16 48 49 9.22
16 30 10.82 30 39 4.24 49 50 5.39

5.3. Optimal waste routing between the transfer stations and the transfer station and processing plants are capable to carry
processing plants the heavy vehicle.
Fig. 8(b) shows the optimal waste transportation path between
In this process, we transfer waste from the transfer station to the transfer stations to the categorized processing plants. The opti-
processing plants through minimum transportation cost. In our mal path reduces transportation cost and time according to Eq.
simulation, we consider 50 transfer stations within the Kolkata (15). In our optimization problem, we compute an optimal path
metropolitan city (shown in Fig. 8(a)). In real life scenario, each for categorize waste transportation between the transfer station
path between the transfer station indicates route between the to processing plants. In this optimization, we consider that vehicle
transfer station and processing plants. Different wastes are trans- must visit at least one transfer station before visiting processing
ferred from the transfer station to categorize processing plant plant. On the other hand, we consider all transfer station must be
using heavy vehicles. The proposed optimization scheme, mini- visited by the vehicle at least one time.
mized transportation cost for categorize waste. After applying the TSP in this section, we got the sequence of
Table 3 shows all possible paths and their distances. We calcu- transfer stations for optimal collection of the solid waste. The
late the optimal path according to these costs. The route between sequence of transfer stations in the case study is: 1 ? 2 ? 4 ? 5
18 S. Das, B.Kr. Bhattacharyya / Waste Management 43 (2015) 9–18

? 6 ? 18 ? 10 ? 7 ? 3 ? 8 ? 9 ? 11 ? 12 ? 13 ? 14 ? 15 ? Eriksson, O., Bisaillon, M., Haraldsson, M., Sundberg, J., 2014. Integrated waste
management as a mean to promote renewable energy. Renew. Energy 61, 38–
17 ? 24 ? 16 ? 25 ? 23 ? 19 ? 20 ? 21 ? 22 ? 27 ? 28 ? 29
42.
? 26 ? 30 ? 39 ? 37 ? 29 ? 31 ? 32 ? 33 ? 34 ? 43 ? 35 ? Gallardo, A., Carlos, M., Peris, M., Colomer, F.J., 2015. Methodology to design a
36 ? 42 ? 38 ? 44 ? 41 ? 46 ? 48 ? 49 ? 47 ? 45 ? 44 ? 40 municipal solid waste pre-collection system. A case study. Waste Manage. 36,
? 50. The total road length in the case study area is 106.26 km and 1–11.
Ionescu, G., Stefani, P., 2014. Environmental assessment of waste transport and
the optimal road length 58.33 km. Therefore, the optimal road treatment: a case study. WIT Trans. Ecol. Environ. 180, 175–185.
length can reduce the number of transport vehicles (heavy). Ana- Ionescu, G., Rada, E.C., Ragazzi, M., Marculescu, C., Badea, A., Apostol, T., 2013.
lysing of this section, we can see that this program marginally Integrated municipal solid waste scenario model using advanced pretreatment
and waste to energy processes. Energy Convers. Manage. 76, 1083–1092.
reduces the transportation cost for segregated solid waste. Jacobsen, R., Buysse, J., Gellynck, X., 2013. Cost comparison between private and
public collection of residual household waste: multiple case studies in the
Flemish region of Belgium. Waste Manage. 33 (1), 3–11.
6. Conclusions
Kanchanabhan, T.E., Abbas Mohaideen, J., Srinivasan, S., Kalyana Sundaram, V.L.,
2011. Optimum municipal solid waste collection using geographical
In this paper, we have faced waste collection and transportation information system (GIS) and vehicle tracking for Pallavapuram municipality.
Waste Manage. Res. 29 (3), 323–339.
problem arising when planning an effective waste management
Khan, D., Samadder, S.R., 2014. Municipal solid waste management using
system. In particular, we have studied the problem of waste collec- geographical information system aided methods: a mini review. Waste
tion in a city where sources are scattered by heterogeneous way. In Manage. Res. 32 (11), 1049–1062.
this paper, we have also studied waste transportation problem for Kim, B.I., Kim, S., Sahoo, S., 2006. Waste collection vehicle routing problem with
time windows. Comput. Oper. Res. 33 (11), 3624–3642.
time and cost effective waste management system design. For this Li, C.-Z., Zhang, Y., Liu, Z.-H., Meng, X., Du, J., 2014. Optimization of MSW collection
problem, in this paper, we have divided the entire waste manage- routing system to reduce fuel consumption and pollutant emissions. Nat.
ment system into three stages. Each stage has been optimized with Environ. Pollut. Technol. 13 (1), 177–184.
Massarutto, A., Carli, A.D., Graffi, M., 2011. Material and energy recovery in
the aid of TSP. We have proposed heuristic solution for optimal integrated waste management systems: a life-cycle costing approach. Waste
waste collection and transportation problem. The proposed Manage. 31 (9–10), 2102–2111.
scheme computes optimal waste collection and transportation Minoglou, M., Komilis, D., 2013. Optimizing the treatment and disposal of municipal
solid wastes using mathematical programming – a case study in a Greek region.
path at each stage. Computational results and real life experiment Resour., Conserv. Recycl. 80 (1), 46–57.
have been shown the effectiveness of the proposed scheme. Results Nuortio, T., Kytojoki, J., Niska, H., Braysy, O., 2006. Improved route planning and
show that the proposed scheme is able to reduce more than 30% of scheduling of waste collection and transport. Expert Syst. Appl. 30 (2), 223–232.
Rabbani, M., Sadri, Sh., Rafiei, H., 2014. Vehicle-routing in a multimodal
the total waste collection path length. This reduction of path length
transportation system used in waste collection. In: IMSS 2014 – 9th
in waste collection and transportation determine consistent mon- International Symposium on Intelligent Manufacturing and Service Systems,
etary saving in the waste management operations. pp. 203–212.
Rada, E.C., 2014. Energy from municipal solid waste. WIT Trans. Ecol. Environ. 190,
945–958.
Acknowledgements Rada, E.C., Grigoriu, M., Ragazzi, M., Fedrizzi, P., 2010. Web oriented technologies
and equipments for MSW collection. In: Proceedings of International
The authors wish to thank the Kolkata as well as Bidhan Nagar Conference Risk Management, Assessment and Mitigation – RIMA’10, pp.
150–153.
municipal corporation administrative authority and their staff for Rada, E.C., Ragazzi, M., Fedrizzi, P., 2013. Web-GIS oriented systems viability for
helping to collect practical data and improve the quality of the municipal solid waste selective collection optimization in developed and
work. transient economies. Waste Manage. 33 (4), 785–792.
Renaud, J., Boctor, F.F., Laporte, G., 2002. Perturbation heuristic for the pickup and
delivery travelling salesman problem. Comput. Oper. Res. 29 (9), 1129–1141.
References Sahoo, S., Kim, S., Kim, B.I., 2005. Routing optimization for waste management.
Interfaces 35 (1), 24–36.
Amponsah, S.K., Salhi, S., 2004. The investigation of a class of capacitated arc routing Wagner, J., Bilitewski, B., 2009. The temporary storage of municipal solid waste –
problems: the collection of garbage in developing countries. Waste Manage. 24 recommendations for a safe operation of interim storage facilities. Waste
(7), 711–721. Manage. 29 (5), 1693–1701.
Aremu, A.S., 2013. In-town tour optimization of conventional mode for municipal Weng, Y.C., Fujiwara, T., 2011. Examining the effectiveness of municipal solid waste
solid waste collection. Nigerian J. Technol. 32 (3), 443–449. management systems: an integrated cost–benefit analysis perspective with a
Bovea, M.D., Ibanez-Fores, V., Gallardo, A., Colomer-Mendoza, F.J., 2010. financial cost modelling in Taiwan. Waste Manage. 31 (6), 1393–1406.
Environmental assessment of alternative municipal solid waste management Wy, J., Kim, B.I., Kim, S., 2013. The rollon–rolloff waste collection vehicle routing
strategies. A Spanish case study. Waste Manage. 30 (11), 2383–2395. problem with time windows. Eur. J. Oper. Res. 224 (3), 466–476.
Consonni, S., Giugliano, M., Massarutto, A., Ragazzi, M., Saccani, C., 2011. Material Xu, S., Li, G., Ou, H., Wu, W., 2015. Research and demonstration of dynamic
and energy recovery in integrated waste management systems: project intelligent logistics system of the collection and transportation process of giant
overview and main results. Waste Manage. 31 (9–10), 2057–2065. municipal garbage. Lect. Notes Electr. Eng. 286, 39–50.
Dumitrescu, I., Ropke, S., Cordeau, J.F., Laporte, G., 2009. The travelling salesman
problem with pickup and delivery: polyhedral results and a branch and cut
algorithm. Math. Programm. 121 (2), 269–305.