Simulation and Modeling 2 (SS 2022)
Vorlesung und Übungen finden als Präsenzveranstaltungen statt. Auf Wunsch ist für die Vorlesung ein Zoomstreaming möglich. Zudem stehen Videos der Vorlesung des Vorjahres auf StudOn zur Verfügung.
|Vorlesung + Übung (2 SWS + 2 SWS)
|2,5 + 5
|Mittwoch, 10:15 – 11:45, Raum H4 UnivIS
|Dienstag, 12:15 – 13:45, Raum 04.158 UnivIS
The class is project-oriented: participants conduct one larger simulation project in a team (3-4 students). The simulation project covers all typical phases including project planning, requirements analysis, data collection, input modeling, conceptual modeling, verification, validation, simulation runs, output analysis, result presentation and documentation. The lecture covers:
- simulation project management, presentation, and documentation techniques
- detailed simulation project case studies,
- advanced topics in simulation such as variance reduction techniques, rare event simulation, parallel and distributed simulation, standards-based simulation, and combination of simulation with testing.
The project teams also present their results in the lectures. The exercises are used for team meetings. Implementations, simulation runs, etc. can be performed on computing facilities of the Computer Science 7 Group with commercial/academic simulation packages (e.g., AnyLogic/OMneT++/ns-3/SimTAny) in reserved computer hours. Possible projects encompass but are not restricted to: elevators, university canteen (Mensa), crossing with traffic lights, gas station, supermarket, hospital emergency reception, flexible manufacturing system, houses or neighborhood with renewable energy systems, battery powered devices. Own project ideas are possible and welcome.
- Project Initialization and Definition Goals: form a team, select a topic. Definition of requirements (what the results should be, not how they are obtained).
- Project Planning Identification of main activities, effort estimation, scheduling (which activity is performed by whom and when).
- Conceptual Model Definition Definition of a conceptual model.
- Model Programming and Data Collection The conceptual model is elaborated and verified in this phase. Data needs to be collected.
- Programming and Input Model Validation Further programming of your model. Construction and validation of input models.
- Programming and Verification Input models are integrated into the system model – a step, which also needs to be verified.
- Validation / Testing Pilot runs serve to validate the model by comparison with the existing system. After validation, experiments are designed, productions runs conducted, and the output analyzed.
- Further Experiments and Animation / Demonstration Further experiments are conducted, a model animation is implemented for presentation purposes. Documentation for the results is to be prepared.
- Project Finalization Simulation results are presented, projects are analyzed. Project reports (20-30 pages) must be submitted.
Ausgewählte ehemalige Projekte
An Energy Demand Model for the Microscopic Simulation of Plug-In Hybrid Vehicles
In this simulation project, an accurate yet computationally inexpensive energy demand model for the microscopic simulation of plug-in hybrid vehicles has been implemented in MATLAB. The simulation model mostly consists of a combustion engine, an electric motor and, most importantly, a realistic engine management model. Furthermore, a pragmatic kinematic model as well as a model for recuperation of energy while breaking are provided. Modularly designed, this simulation model allows for simple replacement and adjustment of each module. As input values, the model only requires the vehicle’s speed and a set of constant predefined parameters. Consequently, this also enables the model to be easily connected with current sophisticated traffic simulators. The correctness of the model regarding the consumption of fuel and electric energy has been validated using data of comprehensive test drives with a real PHEV. The results of this project will be presented by students at the IEEE Transportation Electrification Conference & Expo (ITEC’2016) in June 2016.
Energy Management on Smartphones
With the global growth of the market for smartphones new business ideas and applications are developed continuously. These often utilize the resources of a mobile device to a considerable extent and reach the limits of these. This project focuses on the simulation of an ondemand music service on a modern smartphone. The simulation model includes higher level descriptions of the necessary hardware components’ behavior and their energy consumption. Thereby, the detailed simulation of battery plays a key role in the project. The goal of this simulation study was to find optimal parameters for the users of the examined application to maximize playback time, to improve its battery life and to reduce costly data transmissions. The results of this project has been published at the 17th International GI/ITG Conference on “Measurement, Modelling and Evaluation of Computing Systems” and “Dependability and Fault-Tolerance” (MMB & DFT 2014). The authors have received the “Best Paper Award”. I. Alagöz, C. Löffler, V. Schneider and R. German, Simulating the Energy Management on Smartphones Using Hybrid Modeling Techniques. Proceedings of the 17th International GI/ITG Conference, MMB & DFT 2014, March 2014, p. 207-224. [Paper] [BibTeX] Official PDF Presentation Slides
The project is about simulating the rescue service system in the region of Nürnberg, Fürth and Erlangen. As far as the real world scenario is concerned there is one emergency dispatch center in Nürnberg performing emergency call service management. The corresponding ambulance vehicles are hosted in rescue service stations all over the region. Beside emergency services, as a provision for people with limited physical abilities the rescue service system is responsible for transporting persons from their homes to a physician and back respectively. The major costs of a rescue service system arise from the amount of vehicles in expostulation, in particular these costs are due to personnel costs of the employees and maintenance costs of the vehicles. Accordingly the major goal of the project is to find the minimum number of vehicles necessary to accomplish the arising tasks. A decisive measure in this context is the so-called critical phase, which is the time period where all ambulance vehicles (so-called RTWs) and all patient transport vehicles (so-called KTWs) are in action so that another incoming emergency call cannot be handled until at least one of the vehicles returns. In this regard the duration of the critical phases has to be minimized with the smallest number of vehicles possible. Another aspect when determining the amount of vehicles necessary is the average utilization. In this context two measures are examined. The first one is the average number of vehicles in use. The second one is the percentage of time a vehicle is in action compared to the time it is not used. Report
In this project, the bus line No. 289 in Erlangen from Büchenbach Nord to Waldkrankenhaus is considered. A bus line can be treated as a discrete-event system. For instance, the number of customers on the bus and the travel distances of customers are random variables. Moreover, interarrival times of customers at the bus stop are discrete events. The simulated time is the rush hour, i.e., from 7:00 am to 9:00 am. Although it is possible to simulate the whole day’s operation of the bus line, one of the purposes of this project is to optimize the configuration of the bus line to avoid crowding, but usually crowding does not happen except for the rush hour. The motivation of this project is to model the bus line and obtain the relevant performance measures dependent on an input time schedule. During the rush hour, there are plenty of customers who want to get on the bus. On the one hand, if the time interval of two adjacent busses is too long, considering limited capacity of the bus, crowding will happen. On the other hand, if the time interval is too short, the number of busses will increase, which will generally increase the cost of the bus company. Therefore, an optimal scheme to configure the time schedule needs to be found.
Object of study is the so called Drinks Terminal (in German: Getränke Terminal), an automated retail trade shop for beverages. No service personnel is employed, customers cannot enter the business premises but place their orders at a terminal and are eventually served by crane robots, which also labour as warehousemen. The objective of this work is to build an abstract model of the drinks terminal which should resemble as close as possible the behaviour of the real system. The model is used to simulate how the functional entities of the terminal – the crane robots and conveyor belts – work under certain conditions, like normal load (i.e. number of customers), heavy load and in the case of errors. We assess the system according to the following performance measures: mean time customer spends in queue, mean time between finishing payment and leaving the terminal (which includes the interaction of cranes and conveyor belts to serve the customer), queue lengths averaged for both terminals, utilization of the cranes, customer throughput per hour.
Today’s cash desks in supermarkets are a common bottleneck in our everyday task to collect food in the urban domain. Being the one point where one has to interact with service personnel directly, cash counters usually kick customers out of their shopping flow by making them wait. A cashier waiting for new arrivals is most desirable for the just-finished-shopping customer, whereas most undesirable for the management, for he is someone being paid for doing nothing. So supermarkets (especially large ones) tend to save employment cost by reducing the number of working cash desks to an optimised minimum, which turns out to be a trade-off with customers’ patience. At every cash desk a queue is forming, growing and shrinking dependent on new customers arriving and served customers leaving the place. A supermarket in Erlangen has introduced a mechanism to cheer up waiting customers. The management imposes upon itself a penalty fee for every customer waiting for more than five minutes while not all cash desks are in service. This simulation project investigates the system of waiting people at supermarket cash counters. The correlation between waiting queue length and the number of cash desks in service is the central aspect of this study. The simulation will produce measures to prove that the number of cash counters (servers) affects the mean queue length and waiting time of customers in the queues. The obligatory goal of this study is to inquire about how many cash counters are to be engaged in order to keep the mean waiting time/mean queue length below certain thresholds given a certain amount of traffic. The (simulation-based) computed correlation thus will answer the question, how the supermarket can avoid the penalty fee expenses.
- Prior participation in the course „Simulation and Modeling 1“ is strongly recommended
- Averill Law: Simulation, Modeling and Analysis, 5th Edition, McGraw-Hill, 2014. 3rd Edition available in “Gruppenbibliothek Informatik” in “Handapparat” No. 35.