1.Data model Vs Simulation model in big data
2.Data Modeling
3.Simulation model
4.Preliminaries (correlation & causality)
5.Levels of system analysis
6.Comparison of data modeling and simulation modeling
7.Limitations of data modeling (changed condition)
8.Limitations of data modeling (Unexpected event)
9.Limitations of Simulation modeling
10.Comparison of data modeling and simulation modeling
11.Greenhouse control system
12.Greenhouse control system (Simulation modeling)
13.Greenhouse control system (Data modeling)
14.Greenhouse control system (Results)
1. Data model Vs Simulation model in big data
• Data Modeling
• Simulation Modeling
• Preliminaries
• Comparison of data modeling and simulation modeling
• greenhouse control system
BY : AYA NASRI – RAMI AL-SSADAQA
2. Data Modeling
• Data Mining
• Machine Learning
• It Represents clearly the
correlation between inputs
& outputs
By : Aya Nasri – Rami Al-ssadaqa
3. Simulation model
• Uses physical or operational laws
• It is possible to represent clearly the causality between
inputs & outputs
• It needs prior knowledge of the target system
By : Aya Nasri – Rami Al-ssadaqa
4. Preliminaries (correlation & causality)
• Causation implies correlation, the reverse is not necessarily
true
• Examples:
Car Speed Vs Distance
Quantities of imported oil
vs Chicken consumption
By : Aya Nasri – Rami Al-ssadaqa
5. Levels of system analysis
• Descriptive analysis : explains what has happened in the
system
• Predictive analysis : gives the prediction of what will
happen in the future
• Prescriptive analysis : gives the discernment of how we can
make things happen.
By : Aya Nasri – Rami Al-ssadaqa
7. Comparison of data modeling and simulation modeling
• Limitations of data modeling
• Prediction under the changed condition
• Unexpected events
• Limitations of simulation modeling
By : Aya Nasri – Rami Al-ssadaqa
8. Limitations of data modeling (changed condition)
By : Aya Nasri – Rami Al-ssadaqa
Queuing system example :
• Inputs : inter-arrival time , service time
• Inter-arrive : follows exponential distribution
• Service time : follows normal distribution (μ = 3 , σ = 0.3)
• Output : the turnaround time of past customers
9. Limitations of data modeling (changed condition)
By : Aya Nasri – Rami Al-ssadaqa
Queuing system example :
Very good Results!
What will happen after changing the
normal distribution parameters ??!
10. Limitations of data modeling (changed condition)
By : Aya Nasri – Rami Al-ssadaqa
Queuing system example :
• Change the model parameters to
( μ = 4 , σ = 0.4 )
Very bad !
• Change the problem from
queue(FIFO) to Stack (FILO)
Very bad !
11. Limitations of data modeling (Unexpected event)
By : Aya Nasri – Rami Al-ssadaqa
Average waiting time of passenger in Daejeon example :
• Inputs : (NP) ,(NV)
• (NP) : the number of passengers per hour
• (NV) : the number of vehicles in the city
• Output : average waiting time of passenger
12. Limitations of data modeling (Unexpected event)
By : Aya Nasri – Rami Al-ssadaqa
Average waiting time :
Very good Results!
What will happen after an accident or
the speed limit in some roads??!
13. Limitations of data modeling (Unexpected event)
By : Aya Nasri – Rami Al-ssadaqa
Average waiting time :
After changing roads limit
Very bad !
Why ??
14. Limitations of Simulation modeling
By : Aya Nasri – Rami Al-ssadaqa
• Simulation requires extensive physical and operational
knowledge of a target system in order to be accurate
• if knowledge about the system can be obtained, it’s
preferred to be applied in the prediction
15. Comparison of data modeling and simulation modeling
By : Aya Nasri – Rami Al-ssadaqa
17. Greenhouse control system
• It’s a big data problem , and there is a lot of simulations
solutions to it
• But now we want to use data models with simulation
model to solve it
By : Aya Nasri – Rami Al-ssadaqa