Simulation and Modeling 3.2.pdf

Simulation & Modeling Steps of Modeling Process Part 2 Model Implementation This step is followed by Solving The Model where the conceptual model generated in the previous steps is now have to be implemented as a simulation program. ○ Graphical Environment such as SIMAN or GPS...

Simulation & Modeling Steps of Modeling Process Part 2 Model Implementation This step is followed by Solving The Model where the conceptual model generated in the previous steps is now have to be implemented as a simulation program. ○ Graphical Environment such as SIMAN or GPSS are widely used but C++/Java can be used too ○ Helps a Non-programmer analyst to run it for simulating and analyze ○ Appropriate Hardware configuration is also important for this step (Note: The difference between “Solve The Model” and “Model Implementation” is, Solve the model is the arrangement of simulating the model where Model Implementation is when the model is running) Verify & Interpret Model’s Solution Identify if the model is “Solving the problem right” - Verification and if it is “Solving the right problem” - Validation. ○ Solution might be feasible for a couple of days but proven wrong in long term (e.g. Months, Years) ○ Must be able to predict system performance for wide range of inputs. ○ The comparing should be done with historical data If the model shows any weaknesses, it is advised to cycle back in previous steps and make necessary “Simplification” or “Refinement” which can be a trade-off according to multiple needs. Execution There are two states in this case, Terminating and Steady State. ○ Terminating systems are that runs for a particular time and stops. E.g. Stores, Restaurants etc. ○ Steady states keeps on running and never stops. Such as hospitals, power generators etc. Initial conditions also have profound impact on the models. They determine the startup condition of the simulation. Execution Simulated models are advised to execute several times to obtain different quantitative values for varying random inputs. Initial Conditions may include determining whether the system started in an idle state or a busy state. There must be a set of initial conditions determined. ○ Simulations must be executed several times for each set of condition. ○ For terminating solutions, the system should be run from top to bottom several times for each set of conditions. ○ In case of steady state, data samples should be taken from several subintervals in each run. Output Analysis Analyzing by using appropriate statistical methods. Goal is to determine the the degree of correspondence between the outputs of the model and those of real system. ○ Methods vary depending on terminating or steady-state system ○ Data should be examined after statistical equilibrium is reached for steady- state systems ○ Initial Condition bias can lead to erroneous conclusions Report On The Model Report describes the model in scientific literature so that it can be shared among researchers and laboratories and engineers who can execute it for their use cases. a) Analysis Of The Problem : Clearly explaining the problem and the objectives b) Model Design : Incorporating detail with data, graphs, diagrams, relationships among variables with proper titles and sources divided by columns and axes. Report On The Model c) Model Solution : Describing the techniques for solving the problem and solution with proper equation and information that an audience with minimal expertise can understand. d) Results and Conclusions : Report should include results, interpretations, implications, recommendations and conclusion of the model and future work. Recommendation Finally, the results are to decide what changes to make to the real systems. ○ Changes are judged by their cost association ○ Judgement will be on savings or additional revenue generation by making changes ○ Effect on other systems interacting with the current one The model can supply valuable data when changes are to the system are proposed again. Organization of Program Organization of Program 1. Simulation clock 2. Timer 3. System state variables a. Entity (object) b. Attributes (variables defining object’s properties) c. Events (functions) d. Resources and List Processing 4. Timing routine (to determine the next event, activities and delays) 5. Statistical counters (to decide generation of random events) Organization of Program The simulation begins when the main program invoke the initialization routine, until then the timer is set to zero The system state, statistical counters and event list get initialized After initialization, the control returns to main program when it invokes the timing routine to determine which type of event is most imminent Organization of Program If an event is next to occur, the simulation clock is advanced to the time of occurrence of that event type and control is returned to the main program Then the main program invokes event routine for ensuring that (1) the system state is updated (2) statistical counters are updated and (3) the times of occurrence of future events are generated and added to the event list Generation of random observations from probability distributions is made to determine these future event times using library routine Organization of Program After all processing has been completed, a check is made to determine if the simulation should now be terminated If yes, the report generator is invoked from the main program to compute the estimation of the desired measures of performance and to produce a report If not, control is passed back to main program and the cycle is repeated until the stopping condition is eventually satisfied