Soft Computing Letters 3 (2021) 100008

Contents lists available at ScienceDirect

Soft Computing Letters

journal homepage: www.elsevier.com/locate/socl

Investigation on biomedical waste management of hospitals using cohort

intelligence algorithm
Poorva Agrawal∗, Gagandeep Kaur, Snehal Sagar Kolekar
Department of Computer Science, Symbiosis Institute of Technology, Symbiosis International (Deemed University), Pune 411045, India

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

Keywords: With the innovative development of advanced technology in the field of medical, there is an enlargement in
Biomedical waste management the generation of other problems such as management of biomedical waste. Hazardous waste generated from
Shortest time period hospitals is required to be managed within time and it can be done effectively using some computer science
Total travel distance
technology. In the proposed methodology, Biomedical Waste (BMW) problem is solved with the consideration of
Route optimization
route optimization. Route optimization is important in BMW management because while transporting the BMW
Cohort Intelligence algorithm
from hospital to depot (disposal site) there are many types of risks associated with that route like traffic, vehicle
failure, road accident etc. To avoid the dangerous effects of BMW on humans and environment, it is necessary
to optimize the distance. It can help in promoting healthy and risk free life. This paper addresses the problem
of finding the shortest path using Cohort Intelligence algorithm for BMW management with the consideration of
human risk.

1. Introduction to disposal site, there is risk associated with the transportation. So,
there is a need to have short distance between disposal site and hos-
Improper management of BMW is not only hazardous to environ- pital. The risk related with transportation and collection of BMW is of
ment, but it may also affect the people working in the hospital. It is two types: risk associated with humans and risk associated with the
significantly essential to manage the waste generated within the hospi- environment [2].
tals properly so as to avoid health problems and environmental risks. In this paper, BMW management problem is solved by using newly
BMW includes soil, water, and air pollution, it is mandatory to man- emerging cohort algorithm. Cohort Intelligence (CI) algorithm is evolu-
age the biomedical waste properly as human beings fight with infec- tionary algorithm which shows positive results in various fields. The
tious diseases like HIV, hepatitis etc. The hospital waste management main problem related to BMW management is to provide the safety
involves management of lots of activities, such as collection, transporta- for people while transporting the waste to disposal site. So, there is
tion, treatment, and disposal of wastes. The variety of waste is generated a need to focus on waste management problem using route optimiza-
in the hospital which is hazardous and needs to be transported and dis- tion techniques. These optimization techniques are used to optimize the
posed properly. distance between collection point and the disposal site using vehicle
In last few years, many researchers worked on issues related to routing problem and optimization techniques. The main objectives con-
biomedical waste and they focused on collection, separation, treat- sidered for this problem are: determination of possible routes, study of
ment of biomedical waste in hospitals by using bins. But there is also the risk associated with collection and transportation of disposal and
a need to focus on safe transportation of biomedical waste. In the calculation of the optimized route length.
recent years, many authors worked on network optimization in the The outline of the paper is as follows: Section II discusses about
field of BMW. In 2018, modified Multi-Objective Ant Colony System the literature survey of the different optimization algorithms which
(MOACS) [1] technique was introduced which is facilitated by cluster- are used to solve waste management problems such as municipal solid
ing algorithm to attain safe and optimum routes for Biomedical Waste waste, hazardous waste. Section III describes the proposed idea and the
Complex Transportation (BMWCT) problem. These optimization tech- methodology. Section IV and V discusses about the result and compar-
niques are used to minimize the route, cost related with the trans- ative analysis and the last section discusses the conclusion and future
portation and collection. While transporting the waste from hospital scope of the work.

∗
Corresponding author.
E-mail addresses: poorva.agrawal@sitpune.edu.in (P. Agrawal), gagandeep.kaur@sitpune.edu.in (G. Kaur), snehal.patil@sitpune.edu.in (S.S. Kolekar).

https://doi.org/10.1016/j.socl.2020.100008
Received 20 March 2020; Received in revised form 2 August 2020; Accepted 12 October 2020
2666-2221/© 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)
P. Agrawal, G. Kaur and S.S. Kolekar Soft Computing Letters 3 (2021) 100008

2. Literature review ing problem with time windows constrictions (FT-SMDVRPTW). In this
paper, researchers formulated a solution consisting of a set of directions
This section discusses about the different optimization techniques linked with the trucks, which targets to maximize the overall profit. All
like Ant Colony Optimization (ACO), Genetic Algorithm (GA) etc. These the routes are considered as a series of nominated orders to be served
techniques are used for solving the vehicle routing problem in the in order to preserve the operational constraints. The issue becomes visi-
field of waste management, logistic management, student transporta- ble when automobiles return back. The authors modified the ant colony
tion management etc. There are different types of waste generated from system by integrating it with a robust optimization methodology so as
the society such as hazardous waste, solid waste and biomedical waste. to tackle the full truckload.
This section discusses about various problems that can be solved using Chen et al. [7] examined the location-routing problem with full
the optimization techniques. truckloads (LRPFT), which is the extension of prevailing readings on lo-
cation routing problem to full truckload problem inside the local many
2.1. Background to many raw material supply network. The authors developed a sta-
tistical model which can effectively find the amount and localities of
Kulkarni et al. [2] proposed an evolving technique instigated from amenities and enhance the flows between various types of nodes and
the natural and social propensity of individuals to improve their own ways of trucks. This mathematical model has also reduced the over-
conduct by watching the conduct of other applicants in that abettor all cost and ecological effects. In this paper, authors projected a hy-
called as Cohort Intelligence (CI). This helps in making each individual brid approach called as NSGA-II-TS, which is a blend of multi-objective
advance its behavior and ultimately the complete cohort behavior gets algorithm, Nondominated Sorting Genetic Algorithm (NSGA-II) and a
evolved. The authors tested this competence of the method by resolving recognized heuristic, Tabu Search (TS). The authors authenticated the
an NP-hard combinatorial problem like Knapsack Problem (KP). The computational efficiency of NSGA-II-TS by mathematical outcomes.
results proved that the proposed methodology is extremely focused and Rabbani et al. [8] have built up another scientific plan for the me-
far superior than the existing contemporary methodologies. chanical unsafe waste area directing issues. In this study, researchers
Deshpande et al. [1] introduced modified multi-objective Ant Colony concentrated on minimization of three huge criteria including absolute
system (ACS) technique with the clustering algorithm to obtain minimal cost, complete transportation danger of unsafe waste associated to popu-
and safe routes for Biomedical Waste Complex Transportation (BMWCT) lation coverage, and site hazard. To achieve this, authors implemented
problem. The authors considered the risk related with the transportation two multi- objective evolutionary algorithms: Multi-Objective Particle
and collection of BMW, total scheduling time of vehicles and number of Swarm Optimization (MOPSO) and Non-dominated Sorting Genetic Al-
vehicles as the foremost goals for the selection of routes. In the pro- gorithm (NSGA-II). The authors performed various experimentations on
posed method, clusters of the hospital nodes are formed depending on various problems and found NSGA-II more superior as compared to
late time window related with the hospital node and their remoteness MOPSO algorithm.
from the closest depot. Then the scheduling and optimization of the Buhrkal et al. [9] worked on the Waste Collection Vehicle Routing
routes is achieved using modified Multi-Objective Ant Colony System Problem with Time Windows (WCVRPTW) in a city logistics framework.
(MOACS). The results obtained demonstrated the good performance of The problem is concerned with determining the cost optimum ways for
the proposed system. the trucks carrying the garbage so that all the trash containers are emp-
Dua et al. [3] focused on Multi-Depot Vehicle Routing Problem (MD- tied and the trash is driven to dumping locations though valuing con-
VRP) for the transportation of hazardous materials. The authors concen- sumer time windows and confirming that transporters are given the halts
trated on reducing the total anticipated transportation risk while trans- that the law necessitates. To solve the problem, the researchers came up
porting products made of perilous materials to consumers from several with an adaptive large neighborhood search algorithm and also demon-
depots by developing a fuzzy bi-level programming model. In order to strated the expediency of the algorithm by proving that the algorithm
provide the ideal solutions to the programming model, authors designed can expand the aim of a set of occurrences from the literature and also
4 fuzzy simulation based heuristic algorithms. The algorithms helped in for occurrences given by Danish garbage collection company.
exploring the best approaches, allotting consumers to depots and finding Mathur et al. [10] focused on the basic issues of biomedical waste
the ideal routing answers for all depots and consumers. management. In the paper, authors discussed about inappropriate ad-
Vladimir et al. [4] suggested a Fuzzy Bi-Level programming model ministration of waste created in medicinal services offices, how it im-
for advancing the complete transportation chance while conveying re- pacts the network, human services specialists and the earth. The re-
sults of risky materials to clients from numerous stops. The authors con- searchers explained the classification of BMW such as General Waste,
sidered the several additions of the Vehicle Routing Problem with Time Pathological, Radioactive, Chemical, Sharps, Pharmaceuticals, etc. The
Windows (VRPTW). In this paper, a way to resolve SDVRPTW (extension major sources of generation of BMW such as private hospitals, govern-
of VRPTW)– the variant of the activity permitting distinct merchandises ment hospitals, dispensaries, nursing homes, primary health centers are
distribution to the consumers is also proposed. The two metaheuristic also defined in this paper. The medical waste produced by small hospi-
algorithms: genetic and hybrid are also defined that make use of the tals and clinics can be transferred to the nearby bigger hospitals because
distinctive task- oriented operators and approaches like limited route bigger hospitals have adequate space to adjust the medical waste gener-
inversion, initialization of the original population using Ant Colonies ated from neighboring clinics. So, it helps to optimize the distance and
Method (ACM) etc. the cost associated with the collection and transportation of medical
Ramaekers et al. [5] worked on vehicle routing problem to solve lo- waste generated from small clinics.
gistic problem. The authors used the integrated order picking-vehicle Kacker and Tembhurkar [11] proposed a method to assess and inves-
routing problem (I-OP-VRP) in order to carry out assessment of the ex- tigate the threat of biomedical waste management system for circum-
tra expense of permitting consumers to select a desired delivery time stances prevalent in the countries like India. The data used for develop-
window. Various experiemnts reliant on consumer attributes, operator ing this model was attained via questionnaire survey from investors and
size, time window characteristics were carried out in order to inspect specialists of this field. The authors developed a software tool so as to ap-
the service fee under changing conditions. Grounded on the outcomes proximate the threat values and investigate complete threat connected
of ANOVA, authors found out that extra expense of permitting customers with the biomedical waste management system. This can prove to be
to select a desired delivery time window can be substantially influenced valuable to the organizer, designer, or supervisor managing the biomed-
by inspected factors. ical waste management system to comprehend the threat through this
El Bouyahyiouy et al. [6] suggested an Ant Colony Optimization system, and aid the people to select its proper approaches, plans, and
(ACO) to elucidate the full-truckload selective multi-depot vehicle rout- measures concerning avoidance of loss to people and the atmosphere.
P. Agrawal, G. Kaur and S.S. Kolekar Soft Computing Letters 3 (2021) 100008

collected waste represented. There is a group of nearby depot and a

hospital which is created on the basis of nearby spot. Vehicle depot is
denoted by two nodes 0 and n+1. Every vehicle starts from the depot
node 0, visits a subset of customer nodes and end its route at depot node
that is n+1. N= number of nodes(hospitals), N= {0, 1, …, n, n+1}.

3.2. CI approach

In CI algorithm, consider a cohort of hospitals and depot, every hos-

pital and depot is associated with the route. Hospitals are considered
as a node. Every hospital route is associated with the qualities. In this
problem, qualities are considered as a distance and the risk associated
with environment.

3.3. Distance

Distance is calculated by using Euclidean distance method. Solomon

Fig. 1. Connection between depot and hospitals. benchmark datasets are used for distance calculation. Vehicle starts from
depot which is considered as an origin and visits all the hospitals.
Benavent et al. [12] proposed and compared two mathematical for-
3.4. Risk
mulations to obtain optimum solutions for the Probabilistic Pickup and
Delivery Travelling Salesman Problem assuming that the probability dis-
Kacker and Tembhurkar [11] worked on risk assessment model. Risk
tributions of the requests are known. The first mathematical formula-
associated with improper management of BMW can have impact on hu-
tion proposed is characterized by a polynomial number of variables and
man beings and environment. In this paper, authors defined the different
constraints, while the second formulation is defined by an exponential
causes of risk, different locations and different events for which risks are
number of inequalities. Both the formulations address the service of a set
created. Authors created a model to calculate BMW risk for human be-
of multiple prospective requests. The accomplishment of every request
ings and environment. The value of environment risk considered for this
supports a Bernoulli probability distribution. The problem contemplates
proposed paper is taken from this risk assessment model.
for priori route for picking up and delivering all requests, but a single ve-
hicle following priori route only serves the requests made, thus leading
3.5. Steps to follow in CI
to a posteriori route. The goal is to minimalize the anticipated expense
of the posteriori route. The computational results obtained did not show
The steps to implement CI are shown in Fig. 2:
any supremacy of one formulation over the other, but there were some
instances that were solved by the second formulation only. • Initialize the number of cohorts C, candidates (hospitals), Qualities
Hernandez-Perez et al. [13] introduced a heuristic method to solve of the hospitals Xc , Number of variations t.
the pickup-and-delivery problem. The authors designed routes of least • Consider a vehicle which starts from the depot, visits all the hospitals
expense for a capacitated vehicle transferring goods between a set of and depending on that form three random possible routes.
consumers. Two variants were addressed depending on if pre-emption • Route covers all the hospitals by considering different paths.
is permitted or not. The authors defined a constructive method using • Then calculate the total distance associated with those routes
which the problem is transmuted into a non-split variation first, and • i.e. (F(x)).
then a meta-heuristic algorithm is applied to this variation. As the routes • The probability associated with behavior being followed by ev-
with pre-emption are not generated, so another math- based procedure ery candidate in the cohort is calculated using Probability function
is used at later stages to potentially refine the route. This refinement (Pc ) = Pc 1 +Pc 2
method uses a branch-and-cut algorithm to solve a Mixed Integer Lin-
ear Programming (MILP) formulation. The testing of the complete pro- 3.5.1. Roulette wheel approach
cedure on benchmark instances with upto 500 customers exhibited good • In roulette wheel approach probability Pc of every selected route
performance as compared to existing approaches. According to the au- associated with the depot and hospitals is calculated.
thors, worst-case analysis of the problem and extension of results to • Every candidate expands the sampling interval of every quality
numerous products is challenging for future research. based on whether condition is satisfied. Consider the qualities of
hospital are distance and risk, which are calculated as:
3. Proposed methodology
1
𝑥𝑐𝑛 = ×𝑟 (1)
𝑑𝑛𝑐
This section discusses the methodology to solve the biomedical waste
management problem using Cohort Intelligence algorithm, also referred Quality of route (Xc ):
as an emerging evolutionary Artificial Intelligence algorithm.
1 1 1
𝑋𝐶 = × 𝑟, 𝑐 × 𝑟 ⋯ 𝑐 × 𝑟 (2)
𝑑1𝑐 𝑑2 𝑑𝑁
3.1. Proposed method for biomedical waste management
Calculate objective function F(Xc ) as:
The CI algorithm is used to solve the NP-Hard problem such as { }
F(X𝑐 ) = 𝑥𝑐1 + 𝑥𝑐2 + 𝑥𝑐3 + ⋯ 𝑥𝑐𝑛 (3)
BMW management problem. Results are tested on Solomon benchmark
dataset. The algorithm is tested using the data of 10 hospitals. CI al- • Then apply the roulette wheel approach. In this approach, possible
gorithm is used with the roulette wheel approach to search for the next routes are considered when vehicle is at origin, when vehicle visits
nearby hospital. In Fig. 1, vehicle starts from depot search for the nearby the next hospital and then it searches for nearby hospital, if found
hospital, collects the waste and then again search for nearby hospital then swap (hi , hi+1 ).
from visited hospital. This process repeats till all the hospitals are cov- • In the roulette wheel approach, candidate forms multiple routes for
ered by the vehicle. Then vehicle come back to the depot to dump the t variations.
P. Agrawal, G. Kaur and S.S. Kolekar Soft Computing Letters 3 (2021) 100008

START

Initialize number of routes (candidates) in the

cohort, Qualities of the route, variations t

The probability associated with behavior being

followed by every candidate in the cohort is
calculated

Using roulette wheel approach, all the routes

(candidates) select a route to follow within the
cohort Fig. 3. Optimal route for R101.

Every candidate shrinks/expands the sampling

interval of every quality based on whether
condition is satisfied

Every candidate forms t behavior by sampling the

qualities from within the updated sampling
intervals

Every route follows a t behavior from within its t

behaviors

Every candidate finds the minimal route

Fig. 4. Optimal route for R201.

N
Convergence?

Figure 2: Flowchart of Cohort Intelligence

Accept the current route (cohort behavior)
as final solution

STOP

Fig. 2.

• Depending on minimum distance, find the optimal routes for the

different paths. Fig. 5. Optimal route for RC101.
• After finding different minimal routes, get the one final solution i.e.
optimal route for that network.
• Then depending on the distance from the selected hospital (node) to
the next nearby hospital (r1 , ri+1 ), certain route is selected by the 4. Results and discussion
vehicle using roulette wheel approach.
This chapter deals with results and discussion of the different possi-
The solution has been formed by candidates by swapping and adding ble optimal routes. The routes of ten hospitals to depot were considered
the hospitals one by one. Once the solution is generated then the roulette in this paper. The results provided in the tables shown below are the
wheel approach is applied. There was a slight error while writing that optimal routes between depot to hospital and hospital to hospital. The
has been rectified now. Solomon benchmark data set is used to test the results (Figs. 3–7).
P. Agrawal, G. Kaur and S.S. Kolekar Soft Computing Letters 3 (2021) 100008

Table 4
Possible routes for RC201 dataset.

Data-set Selected routes Total Distance

RC201 1-10-6-5-2-3-8-7-9-4-1 170.83

1-10-9-6-7-8-3-5-4-2-1 157.93
1-7-8-9-10-5-4-6-3-2-1 198.77

Table 5
Route selection by roulette wheel approach for 10 hospitals.

Route t Roulette wheel approach for route selection Total distance

RC201 1 1–10–6–5–2–3–8–7–9–4–1 170.83

2 1–10–9–6–7–8–3–5–4–2–1 157.93
3 1–7–8–9–10–5–4–6–3–2–1 198.77
4 1–10–6–5–2–8–3–7–9–4–1 175.53
5 1–10–6–5–2–8–3–9–7–4–1 182.30
Fig. 6. Optimal route for RC201. 6 1–10–9–6–7–8–5–3–4–2–1 162.62
7 1–7–8–9–10–5–4–6–2–3–1 186.30
R101 1 1–10–9–6–8–7–3–2–5–4–1 263.80
2 1–10–2–6–8–7–9–5–3–4–1 270.63
3 1–7–8–9–10–5–6–3–4–2–1 261.39
4 1–10–9–6–8–3–7–2–5–4–1 271.71
5 1–10–9–6–8–3–7–5–2–4–1 268.06
6 1–10–2–6–8–9–7–5–3–4–1 243.31
R201 1 1–2–4–10–3–5–6–7–9–8–1 216.93
2 1–10–2–6–8–7–3–9–5–4–1 273.62
3 1–7–8–9–10–5–4–6–3–2–1 271.46
4 1–2–4–10–3–5–6–9–7–8–1 229.27
5 1–10–2–6–8–3–7–9–5–4–1 273.03
RC101 1 1–3–4–5–2–6–8–7–9–10–1 157.09
2 1–10–9–6–8–7–3–2–5–4–1 166.32
3 1–7–8–9–10–5–4–6–3–2–1 198.77
4 1–10–9–6–8–7–2–3–5–4–1 172.84
5 1–7–8–9–10–5–4–6–2–3–1 186.30

Table 6
Fig. 7. Graph for amount of waste. Best optimal route for 10 hospitals.

Data-sets Best Optimal Route Total Distance

Table 1 R101 1–10–2–6–8–9–7–5–3–4–1 243.31

Possible routes for R101 dataset. R201 1–2–4–10–3–5–6–7–9–8–1 216.93
RC101 1–3–4–5–2–6–8–7–9–10–1 157.09
Data-set Selected routes Total distance RC201 1–10–9–6–7–8–3–5–4–2–1 157.93

R101 1–10–9–6–8–7–3–2–5–4–1 263.80

1–10–2–6–8–7–9–5–3–4–1 270.63
1–7–8–9–10–5–6–3–4–2–1 261.39
4.2. Roulette wheel approach

4.3. Best optimal route for different datasets

Table 2
Possible routes for R201 dataset.
5. Comparative analysis
Data-set Selected routes Total Distance

R201 1–2–4–10–3–5–6–7–9–8–1 216.93 Many researchers have devoted their contribution to optimize the
1–10–2–6–8–7–3–9–5–4–1 273.62 most important factors like time, distance, cost etc. If the hazardous
1–7–8–9–10–5–4–6–3–2–1 271.46
material generated from the hospital or any other resource is not dis-
posed properly within short time interval, then the people or the envi-
ronment has to face health problem in their life. Many researchers have
Table 3 worked on vehicle routing problem with the well-known optimization
Possible routes for RC101 dataset.
techniques. In this section, the comparative analysis is done for distance
Data-set Selected routes Total Distance between depot and hospital (10 customers)
In 2018, Hu et al. [14] proposed a two-stage algorithm to find the
RC101 1–3–4–5–2–6–8–7–9–10–1 157.09
1–10–9–6–8–7–3–2–5–4–1 166.32 optimal distance for the customer-oriented service strategy. The authors
1–7–8–9–10–5–4–6–3–2–1 198.77 tested the results on different datasets such as R101, R102, RC108,
RC202, etc. To obtain the positive results authors proposed a modified
Adaptive variable neighborhood search (AVNS) method. This method
has been applied on large instances. In this paper, we worked on co-
4.1. Route optimization hort intelligence technique to optimize the route for the transportation
of BMW. The CI has been tested on Solomon benchmark dataset for 10
Possible selected routes for different hospitals on different datasets customers. It shows comparatively good results for the shortest route as
(Tables 1–6): shown in Table 7.
P. Agrawal, G. Kaur and S.S. Kolekar Soft Computing Letters 3 (2021) 100008

Table 7
Comparative analysis.

Difference of Optimal
distance between
Instances Two stage algorithm(MAVNS) Cohort Intelligence MAVNS and CI

Optimal distance Avg time Optimal distance Avg time

R101 287.34 8.72 243.31 7.49 44.03

R102 262.19 6.35 207.56 5.46 54.63
R111 237.40 7.46 207.56 6.41 29.84
R201 259.58 6.42 216.93 5.52 42.65
R202 198.21 6.67 178.76 5.73 19.45
R210 198.21 7.57 192.13 6.51 6.08
R211 198.21 7.34 192.13 6.31 6.08
RC101 239.31 6.53 186.30 5.61 53.01
RC102 203.91 8.39 198.77 7.21 5.14
RC107 202.30 7.29 166.32 6.26 35.98
RC108 202.68 7.37 172.84 6.33 29.84
RC201 212.33 7.14 157.93 6.14 54.4
RC202 203.91 7.65 186.30 6.57 17.61
RC207 204.80 8.23 172.84 7.07 31.96
RC208 202.30 6.82 166.32 5.86 35.98

6. Conclusion and future scope [4] K. Vladimir, M. Ivan, K. Evgeny, B. Tshukin, I. Romanov, Approaches to solve the
vehicle routing problem in the valuables delivery domain, in: Proceedings of the7th
Annual International Conference on Biologically Inspired Cognitive Architectures,
In this proposed methodology, CI has been applied for the ten hospi- BICA, 2016.
tals with one depot (disposal site) to find the optimized route. Compar- [5] K. Ramaekers, A. Caris, S. Moon, T. van Gils, Using an integrated order picking-ve-
ative analysis of this algorithm has been done with two-stage algorithm hicle routing problem to study the impact of delivery time windows in e-commerce,
Procedia Comput. Sci. 88 (2016) 487–492.
i.e. AVNS. These algorithms are tested on Solomon benchmark dataset of [6] K. El Bouyahyiouy, A. Bellabdaoui, An ant colony optimization algorithm for solving
10 customers. Difference between distances calculated by CI and AVNS the full truckload vehicle routing problem with profit, in: Proceedings of IEEE Con-
varies from 5 to 55 km. CI shows the less distance in all the cases as ference on International Colloquium On Logistics And Supply Chain Management
(LOGISTIQUA), 2017.
compared to AVNS algorithm. Risk is also minimized as it is directly
[7] C. Chen, R. Qiu, X. Hu, The location-routing problem with full truckloads in low-
proportional to distance and time. In future, this methodology can be carbon supply chain network designing, Math. Probl. Eng. 2018 (2018) 1–13.
applied to large dataset and real time study as well by considering other [8] M. Rabbani, R. Heidari, H. Farrokhi-Asl, N. Rahimi, Using metaheuristic algorithms
to solve a multi-objective industrial hazardous waste location-routing problem con-
parameters like demand, time, etc. This methodology may also prove to
sidering incompatible waste types, J. Clean. Prod. 170 (2018) 227–241.
be useful to find the optimal route for the student transportations. It may [9] K. Buhrkal, A. Larsen, S. Ropke, in: The waste collection vehicle routing problem
also be applied in medical sciences like ambulance management, supply with time windows in a city logistics context, 39, 2012, pp. 241–254. Proceedings
of blood to hospitals and logistics management scenario for pickup and of the 7th International Conference on City Logistics.
[10] P. Mathur, S. Patan, A.S. Shobhawat, Need of biomedical waste management system
delivery of the products etc. in hospitals-an emerging issue – a review, Curr. World Environ. 7 (2012) 117–124.
[11] P. Kacker, A.R. Tembhurkar, Development of a risk assessment model for an ur-
ban biomedical waste management system, J. Hazard. Toxic Radioact. Waste 19 (3)
References (2015).
[12] E. Benavent, M. Landete, J.-J. Salazar-González, Gregorio Tirado, The probabilis-
[1] R.P Deshpande, A.R. Tembhurkar, Optimal routing of complex transportation sys- tic pickup-and-delivery travelling salesman problem, Expert Syst. Appl. 121 (2019)
tem of biomedical waste with multiple depot and disposal options, Int. J. Environ. 313–323.
Technol. Manag. 21 (2018) 77–90 Nos. 1/2. [13] H. Hernandez-Perez, J.J. Salazar-Gonzalez, B. Santos-Hernandez, Heuristic algo-
[2] A.J. Kulkarni, H. Shabir, Solving 0–1 knapsack problem using cohort intelligence rithm for the split-demand one-commodity pickup-and-delivery travelling salesman
algorithm, Int. J. Mach. Learn. Cybern. 7 (3) (2014) 427–441. problem, Comput. Oper. Res. 97 (2018) 1–17.
[3] J. Dua, X. Li, L. Yu, R. Dan, J. Zhou, Multi-depot vehicle routing problem for haz- [14] C. Hu, J. Lu, X. Liu, G. Zhang, Robust vehicle routing problem with hard time win-
ardous materials transportation: a fuzzy bi-level programming, Inf. Sci. 399 (2017) dows under demand and travel time uncertainty, Comput. Oper. Res. 94 (2018)
201–218. 139–153.