Simulation Analysis for Manufacturing and Software Development

A2. Gain understanding of the reasons for using simulation and get insight into the domains in which it can usefully be applied. (LO2)

B. Subject Specific Intellectual and Research Skills

B1. Solve Monte Carlo and discrete event simulation problems using @Risk and Simul8, respectively

B2. formulate system dynamics problems to solve qualitatively or quantitatively to understand how they are used and how they behave;

B3. experiment using the three different simulation approaches

C. Transferable and Generic Skills

C1. To plan, build and execute with some guidance a small scale simulation study;

C2. To produce a technical and a management (non-technical) report on the simulation study.

DJN, a small avionics firm, is contracted by the military to supply a combat consumable item coded ‘HQ89’. DJN has the capacity to manufacture a total amount of 100 HQ89 units per week. Weekly demand for HQ89 by the military varies. The table below shows the probability distribution of percentage change in demand each week versus the previous week. For example, -30% change indicates the demand in the current week is 30% lower than the previous week.

The management of DJN is considering the following options in order to meet demand for the coming year:

(i) Option1: Maintain current manufacturing levels of 100 guaranteed units of HQ89 per week;

(ii) Option 2: Maintain current manufacturing levels of 100 units of HQ89 PLUS engage a sub-contractor so that the sub-contractor can supply up to 10 guaranteed extra unit of HQ89 per week to DJN. Such extra units would cost £100k;

(iii) Option 3: Speed up the manufacturing process so that DJN can manufacture 110 units of HQ89 a week at a reduced cost of £90k per unit. Unfortunately, speeding up the process reduces the quality of the manufacturing process so that there is now only a 95% chance that each manufactured unit has no defects and will function correctly. This means that there is a 5% chance each of the 110 units will fail. In this situation, the number of non-defective units of HQ89 that DJN will be able to manufacture each week can be predicted by a Binomial probability distribution with n = 110 and p = 0.95. If a unit is defective, then it cannot be repaired or used- it is scrapped. The cost of manufacturing each unit is the same whether the unit is defective or not. The Quality Control is very strict and they can detect the defective units with 100% accuracy.

The option that DJN chooses will not affect DJN’s contractual agreement with the military.  Assume the following:

(i) That the demand at the end of the current year will be 100, so that, for example, if the predicted change in weekly demand is +10%, then the predicted demand in the first week of the coming year will be 100 + 10%*100 = 110.

(ii) That, at the end of the current year, DJN has a residual stock of two HQ89 units so that if, for example, they manufacture 100 HQ89 units in the first week of the coming year, then they will be able to supply 102 HQ89 units in that week.

(iii) That demand figures are WHOLE numbers: If the predicted demand has fractional parts, e.g. 100.2, 100.5 or 100.6, then you must round the number up so that, forexample, 100.2, 100.5 and 100.6 would all equal predicted demand of 101.

(iv) That whichever option DJN selects, the cost of storing the HQ89 units remains the same and therefore irrelevant to the decision.

Your report should be a professional-looking document aimed at lay people who have no expert knowledge of Discrete Event Simulation.

Save the final simul8 model showing the changes you have recommended. This is your final model. Model Submission:

In case we need to check your models, you must submit two Simul8 files: the first should be a base model of the system as it is now. The second should be your final model including the changes that you wish to recommend. Name both using your student number as suffix, for example, as Baseline_123456.s8 and Final_123456.s8.

Task 3 Vensim Model (10%)

A software engineer is managing a new software development project. The project plan suggests that in order to finish the project on schedule, it is necessary to have 35 engineers working on the project. If the number of engineers in the projects falls below this target, the project manager must hire new engineers to the project, a process which takes on average 6 weeks.

The project manager estimates that engineers remain on the project for an average period of 1 year before leaving for another project within the same company or to another organization.

Required:

(a) Create a Vensim model for this system and use it to simulate the system. You must submit a screenshot of your Vensim model and its outputs.

Rate Equations:

• Engineers Hiring rate = Engineers Gap[Number of engineers]/Engineers hiring delay[weeks]

• Engineers gap = Engineers target-Number of Engineers

• Leaving rate[Engineers/week]=Number of Engineers/Average stay period[weeks]

Initial Values

• Engineers hiring delay = 6 weeks

• Average stay period = 52 weeks

• Engineers target = 35

• Initial level of the engineers stock at the beginning of the simulation period = 28

(b) Based on your Vensim model results, discuss how the system will behave in the long-run, i.e. how well the system matches the objectives and, if appropriate, suggest ways in which the system behaviours could be improved in order to achieve the objectives.