See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/330848446

PREDICTION OF SOLID WASTE GENERATION RATE AND DETERMINATION OF

FUTURE WASTE CHARACTERISTICS AT SOUTH-WESTERN REGION OF
BANGLADESH USING ARTIFICIAL NEURAL NETWORK

Conference Paper · February 2017

CITATIONS READS

2 1,495

3 authors:

Md. Ashik Ahmed Md NAZMUL Hossain

North Carolina Agricultural and Technical State University Stamford University Bangladesh
12 PUBLICATIONS   4 CITATIONS    6 PUBLICATIONS   27 CITATIONS   

SEE PROFILE SEE PROFILE

Md. Rafizul Islam

Khulna University of Engineering and Technology
23 PUBLICATIONS   110 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Prediction the thermophysical properties, chemical reactivity, and biological study for Ionic Liquids, natural products, organomatallic crystal and dye. View project

The Food Grains Production AND Fisheries Sector or Fish Production Questionnaire: View project

All content following this page was uploaded by Md. Ashik Ahmed on 04 February 2019.

The user has requested enhancement of the downloaded file.

 PREDICTION OF SOLID WASTE GENERATION RATE AND DETERMINATION
OF FUTURE WASTE CHARACTERISTICS AT SOUTH-WESTERN REGION OF
BANGLADESH USING ARTIFICIAL NEURAL NETWORK

Ashik, M.A.1, Nazmul, M.H.2 and Rafizul, I.M.3

1UG Student, Department of Civil Engineering, Khulna University of Engineering & Technology
(KUET), Bangladesh. Email: ashik1788@gmail.com
2UG Student, Department of Electrical and Electronic Engineering, Khulna University of Engineering &
Technology (KUET), Bangladesh. Email: nazim324@gmail.com
3Associate Professor, Department of Civil Engineering, Khulna University of Engineering &
Technology (KUET), Bangladesh. Email: imrafizul@yahoo.com

Keywords: Solid waste, Waste characterization, ANN, Waste management, South-western region

ABSTRACT

Exact prediction of solid waste generation and its characterization plays a pivotal role in the minimization
and treatment of municipal solid waste management. Nonetheless, it is very tiresome to analyze the
generation of solid waste because of fluctuation of the factors associated with it. Different methods may
have to predict the generation and characterization of solid waste. All the feedbacks were not taken into
account. Here a proposition was given by the use of Artificial Neural Network (ANN) of South-Western
Municipalities and its validation was also perceived. A time series of the generated waste at Khulna
region which have been arranged weekly, from 2008 to 2015, was used for this purpose. The different
inputs were taken as the variable and by conducting the analysis of Regression Analysis interms of
Mean Absolute Error (MAE) as the outputs of ANN were realized.

INTRODUCTION

The prediction of municipal solid waste (MSW) generation plays an important role in a solid waste
management. Yet achieving the anticipated prediction accuracy with regard to the generation trends
facing many fast growing regions is quite challenging. In addition to population growth and migration,
underlying economic development, household size, employment changes, and the impact of waste
recycling would influence the solid waste generation interactively. The development of a reliable model
for predicting the aggregate impact of economic trend, population changes, and recycling impact on
solid waste generation would be a useful advance in the practice of solid waste management.
To plan a solid waste management strategy for a given city or municipality, it is essential to know
the quantity of waste generated and its composition. Various authors in solid waste management have
described the importance of waste characterization. Waste characterization studies provide useful data
on the composition and quantities of solid waste streams (Newenhouse and Schmit, 2000). Bolaane and
Ali (2004) attributed that knowing the waste characteristics is important to waste management policy
making and monitoring. Chung and Poon (2001), stated that data from waste characterization are
essential for waste disposal facilities planning and waste management policy formulation. Solomon
(2011), reported that the waste characterization study on household level provides more detailed,
accurate and crucial information on waste composition and the per capita daily waste generation.
Recognizing the quantity of generated waste is one of the most important factors for operating the solid
waste management system (SWMS), correctly. There are different ways to estimate the waste
generation (WG) rates, which the most prominent of them are load-count analysis, weight-volume
analysis and materials-balance analysis. However, these are the basic methods for estimating the
measure of generated waste, but they have some disadvantages. For example load-count analysis
method determines the rate of collection, not the rate of production. Materials balance analysis method
also suffers from many errors if the source of WG were in a giant size (like a city). On the other part,
traditional methods for estimating the amount of generated solid waste are established, mostly, on the
basis of some elements such as population and social-economic factors of one society and they are
computed according to generation coefficient per person. Since these coefficients change during the
time, so they are useless devices for one dynamics SWMS (Jalil et al., 2008).
Khulna is situated at the south-western part of Bangladesh near the world largest mangrove forest,
Sundarbans shown in Figure 1. The amount of waste has been increasing in recent years, due to the
rapid population growth and improvement of the living standards of the residents. For the rapid growing
of waste it is difficult to predict the amount of generated waste, accurately. For these reasons, employing
new methods and advanced techniques can be useful for computing by means of this dynamic and
non-linear system. These methods mostly consist of some models, classic statistics methods and many
new techniques like time series methods and artificial neural networks (Jalil et al., 2008).
In this study, Artificial Neural Network (ANN) was trained and tested to model weekly waste
generation (WWG) and monthly waste generation (MWG) in South-Western Municipalities of
Bangladesh. The ANN models are basically based on the perceived work of the human brain. The
artificial model of the brain is known as ANN (Sahin et al., 2005). ANNs were first introduced in the
1940s (McCulloch and Pitts, 1943). Interest grew in these tools until the 1960s when Minsky and Papert
showed that networks of any practical size could not be trained effectively (Minsky and Papert, 1969).
It was no until the mid1980s that ANNs once again became popular with the research community when
Rumelhart and McClelland rediscovered a calibration algorithm that could be used to train networks of
sufficient sizes and complexities to be of practical benefit (Rumelhart and McClelland, 1986). Since that
time research into ANNs has expanded and a number of different network types, training algorithms
and tools have evolved. Given sufficient data and complexity, ANNs can be trained to model any
relationship between a series of independent and dependent variables (inputs and outputs to the
network respectively). For this reason, ANNs have been usefully applied to a wide variety of problems
that are difficult to understand, define, and quantify; for example, in finance, medicine, engineering, etc.
Recently, use of ANNs in management of MSW like a proposed model based on ANN to predict rate of
leachate flow rate in place of disposal solid wastes in Istanbul, Turkey (Karaca and Özkaya, 2006),
prediction for energy content of Taiwan MSW using multilayer perceptron neural networks (Shu et al.,
2006). HCl emission characteristics and back propagation neural networks prediction in MSW/coal co-
fired fluidized beds (Chi et al., 2005), recycling strategy and a recyclability assessment model based on
an ANN (Liu et al., 2002) and prediction of heat production from urban solid waste by ANN and
multivariable linear regression in the city of Nanjing, China (Dong et al., 2003), have been become in
current.

Figure 1 Study area: Khulna city (South-Western region of Bangladesh)

MATERIALS AND METHODS

Case study and data:

According to final received reports of KCC, the city has a population of 1.6 million. In latest years, the
rapid growth rate of population increasing to this city has been caused in expanding the WG and as a
result making a problem for the SWMS. According to the CD report, the current actual collected waste
per day is 35% only whereas WG rate is 0.40 Kg per capita per day. In the other hand, the rapid
population growth rate and urbanization cause the significant fluctuations of WG in this city which
consequently results many problems for SWMS. According to Existed reports the amount of generated
waste in Khulna is 500 tons per day, thus offering an appropriate model for estimating the quantity of
generated waste and its fluctuation can be useful for true programming and deciding which is made by
related organizations.

Table 1 Amount of waste generation in Khulna city.

Solid Waste (Tons)

Year
Months 2008 2009 2010 2011 2012 2013 2014 2015
Jan 14780 15030 15340 15700 16300 16650 16900 17200
Feb 13400 13540 13820 14100 14800 15200 15800 16400
Mar 14825 15000 15150 15350 16320 16690 17040 17340
Apr 14650 14265 14650 15410 16300 16700 16900 17250
May 15290 15510 15630 16630 17500 17950 18450 18750
Jun 15150 15470 15670 16100 16400 16700 16950 17260
Jul 15290 15500 15700 16300 16940 17340 17640 17990
Aug 15140 15390 15550 16000 16630 17000 17310 17660
Sep 14650 14850 15050 15500 16100 16460 16810 17100
Oct 15080 15340 15450 15860 16400 16730 17020 17340
Nov 14590 1470 14900 15350 15950 16300 16600 16960
Dec 14980 15200 15340 15700 16180 16470 16770 17070

Table 2 Characteristics of waste in Khulna city

Solid Waste (Tons/Day)

Year Population
Generation Collected
2015 1600000 550 330
2014 1574000 540 324
2013 1530000 530 318
2012 1500000 520 312
2011 1461000 510 306
2010 1422000 500 300
2009 1385028 490 294
2008 1348056 480 288

Physical Composition (%). 1. Food & Vegetable

78.9 2. Paper & Paper Products
100 9.5 3. Polythene & Plastic
3.1
1.14 4. Textile & Wood
1.23 5. Rubber & Leathers
0 6. Metal & Tins
7. Glass & Ceramics
0.23 8. Brick, Concrete & Stones
0.5 0.5 0.1 3.7 9. Dust, Ash, Mud
-100 1.1
10. Egg Cell
-5 0 5 10 15 11. Others

Figure 2 Physical composition of MSW in Khulna city.

 Since having seasonal patterns of generated waste can have an effective role for estimating the
generated waste and its fluctuation in one city (especially in touristy city like Khulna), so a time series
model of WG has been made for predicting the amount of generated waste in Khulna. In this model
weight of waste in t+1 week (Wt+1), is a function of waste quantity in (Wt), t-1 (Wt-1), …, t-10 (Wt-10)
weeks. According to the CD report, the monthly WG in Khulna is shown in Table 1 and characteristics
of SW are presented in Table 2. However, the physical composition of MSW in Khulna city is shown in
Figure 2.

Artificial Neural Network Model:

The neural models are basically based on the perceived work of the human brain. The artificial model
of the brain is known as Artificial Neural Network (ANN) or simply Neural Networks (NN). Neural
Networks have many applications. Generally, however, the ANNs are a cellular information processing
system designed and developed on the basis of the perceived notion of the human brain and its neural
system. Rapid, efficient propagation of electrical and chemical impulses is the distinctive characteristic
of neurons and the nervous system in general. The neurons operate collectively and simultaneously on
most for all data and inputs, which performs as summing and nonlinear mapping junctions. In some
cases they can be considered as threshold units that fire when total input exceeds certain bias level.
Neurons usually operate in parallel and are configured in regular architectures. They are often
organized in layers, and feedback connections both within the layer and toward adjacent layers are
allowed. Strength of each connection is expressed by a numerical value called a weight that can be
updated. Also they are characterized by their time domain behavior, which is often referred as
dynamics. In general, the neuron could be modeled as a nonlinear activated function of which the total
potential inputs into synaptic weights are applied. It is assumed that synapses can impose excitation or
inhibition but not both on the receptive neuron.

Figure 3 Neural network of time series analysis

Figure 4 Close loop view of time series analysis.

Time Series Analysis:

Time series regression is a statistical method for predicting a future response based on the response
history (known as autoregressive dynamics) and the transfer of dynamics from relevant predictors. Time
series regression can help you understand and predict the behavior of dynamic systems from
experimental or observational data. Time series regression is commonly used for modeling and
forecasting of economic, financial, and biological systems. You can start a time series analysis by
building a design matrix (Xt), which can include current and past observations of predictors ordered by
time (t). Then, apply ordinary least squares (OLS) to the multiple linear regression (MLR) model.
 yt = Xtβ + ut

To get an estimate of a linear relationship of the response (yt) to the design matrix. β represents
the linear parameter estimates to be computed and ut represents the residual terms. A neural network
of time series analysis is shown in Figure 3 and a close loop view of time series analysis is shown in
Figure 4. In time series analysis there have 2 input layers, 10 hidden layers and 1 output layer.

RESULTS AND DISCUSSION

The summary of Prediction Performance Statistics Analysis of waste, during the weeks between “2008-
2015”, is given in Table 3. Based on the MAE and values of R, the best neural network is selected; it is
observed that values of R for training and validation come out to be satisfactorily. To achieve the best
ANN structure for estimating generated waste, various structures of time series analysis ANN with three
layers and different number of neurons in hidden layer was investigated. Finally, with consideration on
MAE and R appropriate model was selected. The results of training, testing validation of ANN are given
in Table 3. According to Table 3, the best results were obtained of (2-10-1) structures. These results
are shown in Figures 5 to 7.
The best values of MAE in neural network have been achieved equal to 64.1213 and regression
values for training, test, validation and all, have been achieved equal to 0.92453, 0.91837, 0.93451 and
0.93173, respectively.

Table 3 Results of Training, Test and Validation of ANN

ANN Model Regression (R)

MAE
Structure Training Test Validation All
2-2-1 91.3543 0.93505 0.86161 0.86488 0.91878
2-4-1 52.0435 0.89956 0.90502 0.92296 0.90029
2-6-1 54.6545 0.8997 0.97898 0.96131 0.92116
2-8-1 61.0319 0.924 0.88525 0.88411 0.90855
2-10-1 64.1213 0.92453 0.91837 0.93451 0.93173
2-12-1 64.0112 0.73959 0.5212 0.89307 0.75034
2-16-1 65.0032 0.93489 0.91746 0.91815 0.92844
2-20-1 57.5553 0.90429 0.74016 0.69738 0.85427
2-24-1 62.9976 0.94318 0.5017 0.52496 0.76065
2-28-1 70.1422 0.87125 0.11113 0.8562 0.74118

Figure 5 Performance Plot of Best Neural Network for prediction of waste generation (2-10-1)
Figure 6 Observed and predicted solid waste from ANN Model with structure (2-10-1)

Figure 7 Regression between target vs. training, validation, test and all (2-10-1)

Figure 8 Autocorrelation of Error (2-10-1)

 Figure 8 plot displays the error autocorrelation function of structure 2-10-1. It describes the
prediction errors are related in time. For a perfect prediction model, there should only be one nonzero
value of the autocorrelation function, and it should occur at zero lag (This is the mean square error).
Figure 9 plot shows the correlation between input and error in 2-10-1. This input-error cross-correlation
function illustrates that the errors are correlated with the input sequence. Figure 10 plot shows the error
histogram with 20 bins of 2-10-1 structure. Figure 11 displays the inputs, targets and errors versus time.
It also indicates the time points which were selected for training, testing and validation. All the figures
are generated by MATLAB 7.2.

Figure 9 Correlation between Input and Error (2-10-1)

Figure 10 Error Histogram with 20 Bins (2-10-1)

 Figure 11 Response of output element for Time-Series (2-10-1)

CONCLUSION

The time series analysis of artificial neural network was used by the author for the prediction of weekly
waste generation as well as waste characteristics in Khulna city. In this study, by using ANN model with
one hidden layer and changing the number of neurons of the layer, different models were created and
tested. Finally, according to applied index in this research MAE and R values (training, test, validation
and all), structure with 10 neurons in the hidden layer was selected as the suitable model.

REFERENCES

Bolaane, B. and Ali, M. 2004. Sampling Household Waste at Source: Lessons Learnt in
Gaborone.Waste Management and Research, pp142-148.
Chi, Y., Wen, J.M. and Zhang, D.P. 2005. HCl emission characteristics and BP neural networks
prediction in MSW/coal co-fired fluidized beds. J. Environ. Sci.(China), 17: 699-704.
Chung, S.S. and Poon, C.S. 2001. Characterization of Municipal Solid Waste and its Recyclable
Contents of Guangzhou. Waste Management and Research, pp 473-485.
Dong, C., Jin, B. and Li, D. 2003. Predicting the heating value of MSW with a feed forward neural
network. Waste Manag., 23(2): 103-106.
Jalili, G.Z.M. and Noori, R. 2008. Prediction of municipal solid waste generation by use of artificial neural
network: a case study of Mashdad. Int.J. Environ. Res., 2(1): 13-22.
Karaca, F. and Özkaya, B. 2006. NN-LEAP: A neural network-based model for controlling leachate flow-
rate in a municipal solid waste landfill site. Environ. Model. Software, 21: 1190-1197.
Liu, Z.F., Liu, X.P. and Wang, S.W. 2002. Recycling strategy and a recyclability assessment model
based on an artificial neural network. J. Mater. Proces. Tech., 2002, 129: 500-506.
McCulloch, W.S. and Pitts, W. 1943. A logical calculus of the ideas imminent in nervous activity. Bull.
Math. Biophys., 5: 115-133.
Minsky, M.L. and Papert, S.A. Perceptrons, 1969. MIT Press, Cambridge, MA.
Newenhouse, C.S. and Schmit JT, 2000. Qualitative methods add value to waste characterization
studies. Waste Management and Research pp 105-114.
 Rumelhart, D.E. and McClelland, J.L. 1986. Parallel Distributed Processing: Explorations in the
Microstructures of Cognition, Vol. 1, MIT Press, Cambridge.
Sahin, U., Ucan, O.N., Bayat, C. and Oztorun, N. 2005. Modeling of SO2 distribution in Istanbul using
artificial neural networks. Environ. Model. Assess., 10: 135-142.
Shu, H.Y., Lu, H.C. and Fan, H.J. 2006. Prediction for energy content of Taiwan MSW using multilayer
perceptron neural networks. J. Air and Waste Manag. Assoc., 56: 852–858.
Solomon, A. 2011. The role of households in solid waste management in East Africa Capital Cities.
Environmental Policy. PhD dissertation. University of Wageningeng, Wageningen, The
Netherlands.

View publication stats