Simulacion PEC 2 Solucion Corregida PDF

Preview

Summary

Estudios de Ingeniería en InformáticaSUBJECT: SIMULACIÓN (M1)PEC Num.: 2 Date of proposal: 28 /0 3 /20 Date of delivery: ≤ 27 /0 4 /Observations: The answers will be on this document, keep the original text and take care on the final presentation. It is needed to justify all the answers. The name of...

Description

Estudios de Ingeniería en Informática

SUBJECT: PEC Num.: Date of proposal: Observations:

Evaluation:

SIMULACIÓN (M1.305) 2 ≤ 27/04/20 28/03/20 Date of delivery: • The answers will be on this document, keep the original text and take care on the final presentation. • It is needed to justify all the answers. • The name of the file must be Surname1_Surname2_Name.RTF (o .DOCX o .PDF) All the exercices indicates its weith.

EXERCICES

In this second PEC we are working with the chapters 7-13 of Robinson Book. Follow the “Simulation” guide. Q1 - 15%) (Chapter 9). The time-series graphs below show typical simulation output. For each graph identify the type of model (terminating or non-terminating) and the nature of the simulation output (transient, steady-state, steady-state cycle).

1

Q2 - 20%) (Chapter 10). Ten replications have been performed with a manufacturing model. The mean daily throughput results are shown in Table 10.11 (the warm-up period date have been deleted).

When improving the cycle time of a bottleneck process, new results are obtained, which are shown in Table 10.13. Common random numbers have been used for both sets of experiments. Is the new scenario significantly better than the previous one?

2

Q3 - 25%) (Chapter 12). Carry out some verification and validation tests with a model that represents: A bank is planning its requirements for ATMs (automated teller machines) in a new branch. There are spaces for up to six ATMs, not all of which have to be used. Three types of ATM can be purchased: general ATMs (giving cash, balances, mini statements and PIN change facilities), ATMs for paying money into accounts and ATMs that provide full account statements. The bank has a policy that customers should not wait more than 5 minutes in the majority of cases (generally interpreted as 99%) Q4 - 40%) Read Module 6 of "Simulación con Simio" and do its Activity 4. 1. Develop your own SIMIO models like those in this module. 2. Improve the animation of models using both the standard library and the Google 3D repository. 3. Apply one of the models to your field of interest. 4. Write a brief report summarizing the previous activities. The report should contain: (a) an introduction to the system (including a flowchart), (b) a description of how the model has developed (including images), (c) an experimental section, (d) a discussion of the results, and (e) a conclusion.

3

Solutions Q1 - 15%) (Chapter 9). The time-series graphs below show typical simulation output. For each graph identify the type of model (terminating or non-terminating) and the nature of the simulation output (transient, steady-state, steady-state cycle). Type of model The graph of Figure (d) shows a terminating model because there is no output (or it is null) in the final period. Moreover, in the last periods, the quick fall of the output shows a decrease of the service usually linked to the end of the system process. The graphs of the other figures may show non-terminating models since there is a positive non-stop output. Nature of the simulation output Figures (a) and (b) show steady-state outputs. In the first figure we can see an initial increase of the output (warm-up period) until the simulation arrives to the steady state. On the contrary, Figure (b) may show a steady-state cycle output, with two cycles (the first one larger than the another). Figures (c) and (d) show transient outputs. The mean output of Figure (c) increases. On the other hand, the mean output of Figure (d) first increases and then decreases.

4

Q2 - 20%) (Chapter 10). Ten replications have been performed with a manufacturing model. The mean daily throughput results are shown in Table 10.11 (the warm-up period date have been deleted). When improving the cycle time of a bottleneck process, new results are obtained, which are shown in Table 10.13. Common random numbers have been used for both sets of experiments. Is the new scenario significantly better than the previous one? Table 1 shows the results of the two scenarios according to the 10 replications available for each scenario.

Table 1. Paired-t confidence interval for the two scenarios

The confidence interval (with 95% level of confidence) is the following:

Since all values of the confidence interval are negative, and the objective is to increase the output, we can confirm that the new scenario is significantly better than the previous one.

5

Q3 - 25%) (Chapter 12). Carry out some verification and validation tests with the model of the bank model developed from the case described in Q3 of PEC 1. Next some methods, among others, of verification and validation for the bank model are listed and briefly explained. - Persons with experience in the use of ATMs can be consulted during the development of the conceptual model to validate that the hypothesis and simplifications are right and accurate. - Qualitative observation of the animations of the computer model runs. It can be observed whether the behavior of the queues, use of the ATMs, flow of the clients, etc. to verify that the computer model implements the conceptual model. - If the reliability of the input data is low, a sensitivity analysis can be performed to evaluate the impact of inaccurate data in the outputs of the simulation. - If there are historical data of the system, then a simulation can be run using this data and to validate the output of the model with the real output. In any case, when the system is implemented, real data can be collected to perform the validation. - The computer model is simplified so we can predict the output. For example, a deterministic version of the model using the mean values of the input variables (without variability) is implemented. Since the time between consecutive client arrivals is 80% of the time of the service, in a deterministic model should not be queues. Effectively, the deterministic version of the model implemented in Simio does not generate queues (see Figure 1).

Figure 1. Animation of a deterministic model in Simio

6

Q4 - 40%) Read Module 6 of "Simulación con Simio" and do its Activity 4. It has been developed a model for of the labeling and transport of outgoing boxes of an assembly line. When the boxes are labeled, they are collected at the end of the conveyor belt with a pallet truck to the delivery truck (the boxes are very large, and each truck can only carry a box). Figure 2 illustrates the system.

Figure 2. Labeling and transportation of boxes

The load of the boxes in the pallet truck is modeled as the combination of 2 different entities: The box (member) and the pallet truck (parent). The pallet truck or trucks (the number of pallet trucks is an experimental factor) are generated at the beginning of the simulation. When the pallet truck and the box leaves the system (unload of the box to the delivery truck) a new pallet truck is created at the entrance. Thus, a loop of 20m length is modeled, which corresponds to the activity of carrying the box to the delivery truck and come back for another box. There is a narrow path of 12m length in which only a pallet truck can travel. It is modeled as a two-way road where the pallet trucks must wait to be free to pass. The unload time of the boxes is considered null. The objective of the simulation is to find a configuration that prevents an excessive accumulation of boxes at the entrance of the boxes, since it would affect the suitable behavior of the manufacturing process. Thus, 2 outputs of the model are analyzed: mean times of the boxes in the system and maximum number of boxes waiting at the entrance. 2 experimental factors are considered: performance of the labeling machine and the number of pallet trucks. Let suppose that in the market there are 3 types of labeling machines with different labeling times: Less labeling time, higher cost. The basis costs is C. Regarding the number of pallet trucks, it is contemplated the possibility of acquiring 1 to 3. The cost of each pallet truck is also C. Table 2 shows the possible values of the experimental factors and their associated cost. Labeling machine Labeling time Cost (minutes) U[0.2, 0.25] 3C U[0.25, 0.4] 2C U[0.4, 0.55] C

Pallet truck Units Cost 1 2 3

C 2C 3C

Table 2. Experimental factors and their economic cost

7

Thus, 9 scenarios are considered. For each scenario, 50 replications are calculated. The simulation time is 16 hours (2 shits of 8 hours each one). The obtained results are shown in Figure 3.

Figure 3. Results of the simulation

We can see that only 4 scenarios avoid and excessive accumulation of the boxes. The value of the column “Accumulation” is the mean maximum number of boxes in the queue. Thus, the scenarios with only 1 pallet truck or the cheapest (slowest) machine is not advisable. On the other hand, as it is expected, the scenario with the fastest machine and 3 pallet trucks return the lowest time, on average, of the boxes in the system. However, if we take into account the economic cost, the scenario with 2 pallet truck and the medium time machine returns similar results. Therefore, this scenario is recommended to be implemented. Figure 4 shows an screenshot of a 3D animation of this scenario modeled in Simio.

Figure 4. Animation in Simio of the labeling and transportation of boxes

8...