Lecture 7: DES Flashcards
What are the two types of Systems?
- Discrete
- system where status variable change instantenously at discrete intervals of time
- > IT systems, manufacturing system, traffic regulation systems - Continous
- status variables change continously during time of simulation
- > flow of fluids heat
What is the difference between analytical and experimental models?
→ simple systems can u can use mathemtical models (algebral, calculus, or linked to probability)
→ if its too complex (to many variables having interdependancies) u need experemental models throught simulation
Explani deterministic and stochatic models
- Deterministic
- no uncertainty in simulated phenomena (e.g. customers arriving at regular times) - Stochastics
- correlation between variables and probabilities
Types of Simulations
- Statis vs Dynamic
→ does time play a role? (do some parameters change in generell) - Continous vs Discrete
→ do changes take place continously or instantenously at certain points - Deterministic vs Stochastic
-is there randomness in the inputs or operations of the model (how likely is it for a mashine to go out?)
What is SImulation
→ Simulation is the imitation of the operation of a real-world process or system over time.
→ Computer simulation is a methode where experiments are conducted on a computer in order to tecreate over time the functioning of a system
→ experimental tequnique allows to predict the behaviour of a system under certain conditions that most likely will manifest while the system
→
Whats the function of Simulation?
→ recreate behaviour of a system over an extended period of time, in a shorter time
input X → model → output Y
How does Simulation supports in the industrial sector?
- Desgin of new systems/solution (e.g. factory)
- improvment of actual systems/solutions
- Performance analysis (e.g. productivity, bottlenecks, resources saturation)
- WHAT-IF analysis (new order from customer into system)
- Training
What are Discrete event systems? + Examples
→ dynamic systems (over time) which update their values only at discrete number of points
- no continously changing states are of importance
Examples: call center, gas station, restaurant factory, warehouse
(restaurant: customers enter at certain time, they sit at certain etc.)
What is the Approach of DES
- focus on times between events
- with full knowledge we can update the system only at this specific time point
- full info about history, cause nothing happened between t1 and t2
What is an Entity of DES?
-thing with distinct and temporary existence that passes through a system (e.g a piece that should be assemblet, client to the bank office)
→ created to simulate a process and then destroyed
What is a Resource of DES?
thing with active and permanent existence that provides a service to entities (e.g. machine tool, conveyor, a banker, …)
What is a queue of DES?
a thing with temporary and passive existence that is formed by entities which are waiting for a resource (that is not free at that time).
→ always there, but can be empty or full, so temporary
What are Variable and Attributes of DES?
data and information about entities or resources (e.g. warehouse capacity, cycle time).
What is the Status of DES?
Description of all the values attributes of an entity or a resource, in a specific instant of time.
What is an event in DES?
Instant of time in which the model status is changing (change of status of entity or resource).
What are Activities in DES?
Operations done between two events (started and ended)
→ has a start and end time, so time between events
What are the Simulation Phases?
- Real System (Problem)
- Conceptual Model
- Computational model
- Solution
What is the first Phase of the simulation about?
→ describe the Real Sysmte + Problem definition
-identifiy objectives and constraints (what do you want to simulate?)
- define PJ plan:
1. choose the differen simulations sceanario
2. PJ time (how long simulation takes)
3. resources, team, different roles (customers has to give data)
Example:
- want to measure the production time
- constraint: 5 plants available for 8hours a day
What is the second phase of the simulation process?
→ Conceptual Model - detail and accuracy of the model
- Variables (e.g. production plant design or size)
- exogenous (external cause, process time)
- decisional (machines numbers)
- endogenous (internal cause, e.g. saturation rate) - Constraints
- Performance Measurements
- endogneous variables
- choose the objective functions
→ the more detailed, the more accuracy is reached. BUT with too many details the accuracy shrinks
What is the importance of Data colleting in simulating?
-sometimes quite hardworking (digitalization solved it)
-Data quality influences model quality (GIGO effect = garbage in, garbage out)
→ bad data, bad result
- if data is missing, u should adjust the model to keep accuracy high
- not enought data lead to statistical distributions
-
What Simulation Software can u use?
Generel-Purpose Languages
- procedural (Fortran, C, Pascal, Basic)
- object-oriented (C++, Objective C, Java)
Simulation Languages
-Siman, Slam, Modism
Simluatios
-simulation packages (e.g. PlantSimulation, Arena, Simul8, Matlab-Simulink)
What are some criteria are there for software?
- Flexibility
– Ease of development
– Model execution speed
– Animation
– Output visualization
– Statistics
– Interface with other data
– Reusability
– Customer support
– Cost
→ has to be able to instantiate the different parts of the sysmem (enitiies, resoucres, queues, ACs) and describe different behavoiur of system (status, events, ACs)
Was ist Model Assessment and Validation?
→ Assessment: accuracy, completeness and consistency of the simulation model respect to conceptual model (Did you do the things RIGHT?)
→ no bugs or problems in the code
→ Validation: accuracy, completeness and cosistency of the simulation model respect to real system
(Did you do the RIGHT things?)
→ solution fits the charateristics of the real system. Is performance in the model = performance real world)
What is the digital twin paradigm?
→ the simulation model is a Virtual replica synchronized with the field
- Digital Model → is the simulation model core (just manual data flow)
- Digital shadow → is the simulaion model that receives data and info from the system (automatic flow from physical to digital object)
- Digital Twin → is a digital shadow that can also feedback the decision and optimizations on the field (automatic data flow)
