Ateneo - UPO

The course provides motivation for performing simulation studies, illustrates the main concepts and methods for building simulation models, executing them and analysing the output. This is done presenting several application examples and proposing hands-on exercises using open source simulation environments. The basic background notions on probability and statistics needed to perform the analysis of the simulation output are reviewed in the course.

Text Book (Open Access): Manuel D. Rosetti, Simulation Modeling and Arena.
https://rossetti.github.io/RossettiArenaBook/
Books for consultation:
J. Banks, J. S. Carson II, B. L. Nelson, David M. Nicol, Discrete-Event System Simulation , Fifth edition, Pearson Education 2010
Simulation Modeling and Analysis, 5th Edition, A. Law, Mc Graw Hill, 2015

This is an introductory course on Discrete Event Simulation. It focuses on the purpose and possible applications of simulation. The concepts of dynamic discrete model, its state and the operational laws defining the model dynamics are first introduced. The goal of a model-based simulation study is then defined providing examples of possible application domains and of the properties (measures) of interest that can be investigated through such study. The problem of choosing the appropriate abstraction level as a function of the desired results and of the available information on the system under study is also discussed.
The difference between deterministic and stochastic models is presented; workload characterization, random numbers and random variate generation as well as statistical analysis of simulation output are introduced as relevant aspects to be considered when performing a simulation study through stochastic models.
A few tools for the design and simulation of discrete dynamic stochastic models are presented highlighting similarities and differences. Getting acquainted with one tool and going through the main steps of a simulation study on a simple but realistic project is also a goal of the course.

Basic mathematics, statistics and programming notions and skills

This is a blended course, with a first on-line part and a final residential part.
In the on-line part: On-line synchronous classes (use of slides and student engagement tool Wooclap), self study on material published on-line, on-line self-assessment quizzes, exercises using open source software tools, on-line support for lab activities. Use of diverse on-line tools for communication and collaborative learning.
In the final residential part of the course: Teamwork on a project. The result of the project work will be presented to all course participants at the end of the residential part of the course.

Blended course with an on-line part (13 weeks, part-time) and an intensive residential part (1 week). Final project to be developed in groups.
The DIR (Moodle) platform will be used as a shared point where all the material and links to all on-line communication tools and open source software tools will be available for the participants. Synchronous classes will be given using a video-conferencing tool (e.g., Meet, Zoom, …).
On-line support for laboratory activities will be performed through Mariotel https://tel.marionnet.org/info.php

The assessment will be based on three aspects: (1) the quality of artifacts produced along the whole course (30%), (2) the quality of interactions, in particular concerning on-line discussion and group work collaborative activities (10%), (3) the quality of the final project work (in group) developed mainly during the residential intensive week and illustrated in a final presentation (60%).

Motivation for developing simulation studies, with examples in diverse application fields.
Discrete Event Simulation: definition of model state and operational rules for state evolution in (discrete) time. Definition of performance/reliability measures. Manual simulation of simple examples.
Discrete Event Simulation: basic Next Event Time Advance algorithm (Future Event List, Event handling methods, Measures computation).
Simulation model implementation tasks are proposed, to be completed through general purpose programming languages or within application oriented open source simulation environments (e.g. JaamSim and Omnet++).
Review of basic probability and statistics notions. (Pseudo)Random numbers and Random variates generation methods.
Experiments with stochastic models: observing input and output variability.
Fitting distributions on input data. Simulation output analysis (of stochastic models): computing confidence intervals on multiple run output. Transient versus steady state analysis of measures.
Presentation of a few case studies through guided experiments on a few case studies with different software tools.
Final project assignment concerning a variation on the theme of one case study proposed in the previous phases: this will be developed within groups (3-4 students ), and comprises the production of a complete simulation study, of a report describing it and a presentation to be discussed with all participants.

After the course the participants will be able to:
Define the purpose of a simulation study and provide examples of possible applications.
Define what a dynamic discrete simulation model is; list the elements that define a model; provide examples of models in different application domains.
Build a model using a modeling and simulation tool and execute it.
Compare the characteristics of different simulation tools.
Compare models at different abstraction levels and discuss what properties or measures can be derived by simulating such models.
Define the purpose and the characteristics of a random number/variate generation method and its role in the execution of a stochastic simulation model.
Derive the parameters of a stochastic simulation model from data gathered on a real system: in particular apply probability distribution fitting methods from datasets.
Explain why and how statistical analysis methods should be applied to the results of (multiple executions of) a stochastic model simulation esperiments..
Apply all the above abilities to a realistic case study. Perform what-if analysis to observe and explain how the simulation results change when considering variations of the model (different configurations or parameter values).
Finally along the course the participants will develop soft skills such as team work, communication skills (both technology mediated and in-presence) and the ability to expand their knowledge on the subject through autonomous learning.