An Examination of Big Data's Influence on Simulation Modeling

This report delves into the significant impact of big data on simulation modeling, exploring the opportunities and challenges it presents. It examines the limitations of data modeling, such as its inability to handle uncertain events and its dependence on pre-defined concepts, in contrast to the flexibility of simulation modeling. The report provides a detailed comparison between data modeling and simulation modeling, highlighting their respective strengths and weaknesses. It discusses the importance of exploratory research and the integration of intensive data and computation. The report also emphasizes the need for efficient data storage, analysis, and management, particularly in the context of big data's dynamic features and heterogeneous datasets. Finally, it concludes that by combining the advantages of both data modeling and simulation modeling, users and organizations can create powerful models for big data analytics, particularly when dealing with complex problems where direct analytical approaches are insufficient.

Running head: BIG DATA ON SIMULATION MODELLING
BIG DATA ON SIMULATION MODELLING
Name of the Student
Name of the university
Author note

1BIG DATA ON SIMULATION MODELLING
Introduction
Big Data era has brought opportunities, confusions and challenges for the field of
simulation and modelling related to big data. Novel research ideas and processes have been
brought by big data. It consists of many features like big data network for object analysis,
imprecision and variety of data. It emphasizes on detailed analysis and analyzing indirectly,
hence it makes big data paradigm of new research. The experts who support fourth paradigm
have the opinion that the paradigm of scientific research which is data intensive has a
considerable impact on simulation (Becker, 2016). Some experts define paradigm transition
as a collection of research methods and processes that is replaced by a second set of processes
and big data lacks support on the study of systematic theory, application practice and
methods which are technical in nature. In this paper, the effect of big data on simulation
modelling will be discussed.
Data modelling limitations
Big data is unable to become a research paradigm independently. It has brought great
challenges to simulation in terms of facing and dealing with numerous problems. Simulation
theory is the first problem and simulation paradigm based on the model is dependent on
causality, predefining many concepts like status, boundary, target, entity, attribute and
constraints. Meeting the demands of big data processing on the social activities of people on
the internet is difficult and also solving the problem in disorganized world. Modelling method
is the second problem and in this the provision in big data to use data model in a new way
facilitates the data model to be aware of the problems and provide a solution to the problems
which are highly complex and having numerous computation (Chen & Venkatachalam,
2017). This makes it difficult to feasibly model and design new model types. Science and
engineering of simulation is the last trouble. Exploratory research is important to be carried

2BIG DATA ON SIMULATION MODELLING
out with the combination of science and simulation engineering for developing a virtual
environment for the realization of the integration of intensive data and computation. This also
involves the realization of casual law that is unorganized and can be seen in complex
systems.
Data modelling has weaknesses and it can be described by the help of predictions of
traffic simulation and queuing method. If it is assumed that a queuing system exists in the
real world and a turnaround time needs to be predicted which means the time user remains in
the system. There are three types of data which make the data-sets and those are as follows:
turnaround time of previous customers, arrival time and service time (Fox, Qiu, Jha,
Ekanayake & Kamburugamuve, 2015). From the perspective of data-modelling, constructing
a static map is of input to output which uses techniques of machine learning considered as the
most important method.
Data model system is unable to cope up with the events that are uncertain. So this is a
limitation of data model, however a real system does not have this drawback. During the
occurrence of the event, the data model corresponding to the data set is unable to predict the
outcome of the uncertain events accurately.
On the other side, simulation model is independent of data set and reasonable results
can be made. For showing the drawbacks, data model has described vehicle allocation system
(VAS). The objective of VAS is finding the average time that a passenger waits. ANN is used
for constructing the data model by utilizing the VAS data set. VAS has a simulation model
which is agent based and utilizes DEVS. Simulation model environment is set as a graph
which depicts the real traffic scenario (Kecskemeti et al., 2017). The simulation as well as the
data model produces accurate results of the events, but both produce varied results when
uncertain events occur. In a case where few roads experience traffic jams due to accident.

3BIG DATA ON SIMULATION MODELLING
Here the model which predicts proper results is the simulation model. So the simulation
model which includes VAS achieves reflection by modification of certain parameters at a
cost which is much lower, whereas the data model excludes uncertain events but produces a
matching prediction to the data set given.
Adding to that when the policies of some roads like speed limit is altered, the data
model needs a modified data set of the environment for predicting the results precisely.
Heavy cost is incurred due to the acquirement of new data set (Kim, Kang, Choi & Kim,
2017). Though the data set is considered as unique event, it is difficult to receive the real
system data. On the other hand in simulation reasonable outcomes can be presented by
modifying certain parameters at lower cost.
Limitations of simulation modelling
It can be observed that simulation model is preferred. There may be some situations
where very less or nothing is known about the system. Extensive knowledge in terms of
operational and physical features of a system for able to be precise. In such situation data
modelling approach should be acquired, if system knowledge can be gained. Moreover
simulation model needs predictions and protocols about the environment whereas it is not the
same for data model (Marshall et al., 2016). In a scenario when in automotive engine
modelling of its controller is executed, the validity of the controller can be for a low
operational range since assumptions are included which are ideal in behaviour and various
model parameters are determined by utilization of experimental data having a steady-state.
This is one drawback of simulation modelling.
Data modelling and simulation modelling comparison
When it is seen from the modelling structure perspective, input variable is associated
with output variable in form of static map in data modelling. This is the reason why without

4BIG DATA ON SIMULATION MODELLING
any kind of knowledge, a modeller can design the data model. Alternatively a dynamic map
to output variable to the input and state. Thus for this approach the knowledge of target
system is required. The previous approach has been used in the studies of computer science
research like machine learning and data mining (Tolk, 2015). Machine learning deals with
research and focuses on causality that is operational and physical regulations. Analysis level
is the important aspect in the modelling approaches. It is known that approach of data
modelling is created on information that is learned and it facilitates for analyzing in low level
like defining the past and predicting the days to come. System is represented as simple static
model and this model execution is operational in real time. On the contrary the approach of
simulation modelling depends on knowledge and analysis which is of higher level. The future
prediction is enabled under varied situations while on training. This difference permits the
approach of simulation modelling for conducting a system analysis with the systems that are
non-existing and abnormal, for instance a unique incident or new design and this is not
possible in the approach of data modelling (Vitolo et al., 2015). Since the model is highly
complex the decision time is longer and produces difficulty in the execution in real time
scenarios.
Modelling as well as Simulation is considered as important tools in the field of
science and engineering. These tools are made use of for predicting and analyzing complex
systems as well as other natural phenomena. Modelling addresses complexities as it raises the
level of abstraction aiming at representing the domain at hand in a proper way. This in turn
has been the reason behind the complicated trade-off in between accuracy and efficiency as
well. System properties can be found out by simulation of behaviour through the abstract
representation of the same. Big data’s dynamic features are far beyond the reach of the
existing modelling and simulation tools. Datasets are heterogeneous in nature that means
these data sets are developed by distinct sources. The datasets are large in size having high in

5BIG DATA ON SIMULATION MODELLING
and out rates. Big data accessibility and capability to efficiently combine data together has a
lot of value. On a daily basis variety of electronic data is created and this is the case after the
inception of Internet. The sources of raw data can be different sources that include mobile
devices, instruments, computer files and sensors and many more. The challenge is to store
this data in a proper way. The big data problem requires that the servers are improved along
with network infrastructure so that efficient analysis as well as interpretation of the data via
the on-hand data management applications. The major challenge in case of big data
modelling and simulation is defining framework that includes intelligent management along
with communication, mapping algorithms and various protocols. Energy-awareness is
considered as a vital aspect of big data computation along with simulation (Wassan, 2015).
The main goal is to bring in reductions in the existing gap between capacity provided by the
distributed computing environments as well as the application requirements. This is required
especially during the low workload time. There have been continuous efforts made to come
up with efficient scheduling of various tasks along with balancing of loads. Both data
modelling and simulation have certain limitations. Like in case of data modelling, the
limitation is that it cannot cope up with the unexpected events.
Conclusion
In today’s world, volumes of informations need to be stored as well as processed and
this is because of the rapidly developing information technology. In this paper the connection
in between big data and simulation modeling can be understood. Big data is an emerging
paradigm and is made use of by the researchers for prediction of target systems in different
research fields. In order to do the same, a specific model is required and there comes the data
modeling and simulation modeling. Data and simulation modeling both have their advantages
as well as various limitations and thus it can be concluded that when these models are made
use of together, they can help the users or organizations as a whole. The merits of both these

6BIG DATA ON SIMULATION MODELLING
models can be combined together so that they form a strong model and help in big data
analytics. Simulation is made use of in case where the problems’ complexity makes applying
direct analytical approaches impossible. Thus this is applicable mainly to the problems
related to big data. Modeling and simulation together have been able to address various
problems that include scheduling tasks in different environments such as distributed as well
as heterogeneous. Simulations have been found useful in handling the complexities of big
data analytics and as per the recent trends it can be said that these models are being used of to
help the programmers establish control over the hardware performance in a better way than
before.

7BIG DATA ON SIMULATION MODELLING
References
Becker, T. (2016). Big data usage. In New horizons for a data-driven economy (pp. 143-165).
Springer, Cham.
Chen, S. H., & Venkatachalam, R. (2017). Agent-based modelling as a foundation for big
data. Journal of Economic Methodology, 24(4), 362-383.
Fox, G., Qiu, J., Jha, S., Ekanayake, S., & Kamburugamuve, S. (2015). Big data, simulations
and HPC convergence. In Big Data Benchmarking (pp. 3-17). Springer, Cham.
Kecskemeti, G., Casale, G., Jha, D. N., Lyon, J., & Ranjan, R. (2017). Modelling and
simulation challenges in internet of things. IEEE cloud computing, 4(1), 62-69.
Kim, B. S., Kang, B. G., Choi, S. H., & Kim, T. G. (2017). Data modeling versus simulation
modeling in the big data era: case study of a greenhouse control
system. Simulation, 93(7), 579-594.
Marshall, D. A., Burgos-Liz, L., Pasupathy, K. S., Padula, W. V., IJzerman, M. J., Wong, P.
K., ... & Osgood, N. D. (2016). Transforming healthcare delivery: integrating
dynamic simulation modelling and big data in health economics and outcomes
research. PharmacoEconomics, 34(2), 115-126.
Tolk, A. (2015, July). The next generation of modeling & simulation: integrating big data and
deep learning. In Proceedings of the conference on summer computer simulation (pp.
1-8).
Vitolo, C., Elkhatib, Y., Reusser, D., Macleod, C. J., & Buytaert, W. (2015). Web
technologies for environmental Big Data. Environmental Modelling & Software, 63,
185-198.

8BIG DATA ON SIMULATION MODELLING
Wassan, J. T. (2015). Discovering big data modelling for educational world. Procedia-Social
and Behavioral Sciences, 176, 642-649.