Reference no: EM132513257
RSK80007 Quantified Risk Modelling - Swinburne University
Part 1. Submit the risk modelling spreadsheet which you have used to replicate the results of the two case studies provided in this module. This spreadsheet must have these features:
a. Text fields used to describe the risk being modelled
b. Text fields used to define data entry requirements for each of the Risk parameters that are used in the creation of the model, eg. What makes up the Consequences? A space for estimating both the LLC and the LWC, etc.
c. Text fields or comments used to describe/explain the origin of numerical estimates included in numerical entries
d. Comments to make it clear how probability has been estimated and to what the probability applies
e. Text fields used to define data entry requirements for all parameters relevant to cost benefit analysis, including both data input and calculated value fields
f. Text fields used to make the meaning (for probability) or units (all others) of all numerical entries explicit
g. A separate sheet that reads needed values from the above for the purpose of drawing a risk diagram with log scales and allowing the total risk to be summed from the relevant data points along the risk diagram.
h. Numerical fields in which to enter values and/or in which to place values calculated from data entries for both the risk model and the cost benefit analysis - the latter reading values from the second (risk diagram) sheet
Part 2. Monte Carlo analysis (MCA, more accurately Simulation, as many of you pointed out)
First let's get terminology translated. The use of terms such as Outcome, Risk, Hazard, analysis, Risk Assessment etc. are used in various ways in the documents you searched for. In this Unit, Outcome has a particular meaning, as does Risk. I have not used the term Hazard, but the text does suggest a sensible definition of it. Analysis in the context of risk as defined has a very specific meaning - see the text - as does Risk Assessment
Risk - is not the same as chance or probability or likelihood. In the sources you used this distinction was not made. I don't want you to confuse this colloquial use of the term with the defined meaning of the term in this Unit.
MCA does not: Analyse risk; do a hazard investigation; predict anything; show possible outcomes
MCA has nothing to do with the notion of risk tolerability and it is not a risk assessment tool.
MCA can be used in any calculation with any application where input variables are of uncertain value.
Second - I am looking for your own words. I am not interested in loads of text pasted from another source. However, this is what most of you did. I suspect I could charge maybe three quarters of you with plagiarism. I have not pursued this, but it gives me a poor impression of your academic honesty or maturity.
Third - I was expecting each of the questions below to be responded to explicitly. Many or most of you, wrote much the same complex gobbledygook for each.
a. In your own words, describe how Monte Carlo analysis works (approx 500 words)
Describing how it works is not the same as describing what it is or where it is used. Examples, where given, should be relevant to QRM, not to project scheduling, the price of houses, flip-ping coins, investment decisions etc.
Here's the simple plain English version. Monte Carlo simulation replaces single point estimates of input parameters in any calculation with a range of possible values, from a perceived minimum to perceived maximum. This is applicable anywhere there is uncertainty in the input value. The fact that some numbers in the provided range are judged more likely to be correct than others can be shown by a probability distribution, for example a Normal distribution or any of a number of statistical distributions. When a single calculation is done, a single value of the input parameter is chosen within the allowed range. When multiple calculations are done the number of times values are chosen within the allowed range follows the distribution that has been selected. Eg, in the case of a Normal distribution, most of the numbers chosen will be in the central area, with few chosen in the outlying areas. The process of choosing a value for the calculation is based on a random number generator whose output is biased by the probability distribution selected. The result of the numerous calculations is an output that reflects the input possibilities. The output will be a result of more than one inputs and so the output will not necessarily have the same distribution of values as any one input parameter. (Doing this on a computer makes it easy, but it can also be done by hand)
b. In your own words, explain your understanding of how Monte Carlo analysis can benefit a quantified risk model (approx 500 words)
Note the question excludes explanations related to project schedules. It is the benefit that is of interest, not a rehash of how it works or where or why.
A QRM inherently makes use of uncertain values, often ones that have been guessed using the best judgement possible. Using MCA the sense that we have of how uncertain our input values are can be reflected in the range and distribution we choose. In this way the result of the calculation is not necessarily a narrow line on the Risk Diagram but a wide and fuzzy one. The result clearly indicates our uncertainty whereas a single number result makes it appear that we are more confident than we really are. Whether we do it this way or not , we can still see how sensitive the model is to any one input value by just using one input value from the low end of the range and one at the high end of the range.
c. Practically, how does a typical Monte Carlo analysis spreadsheet insert contribute to the management of uncertainty in the values used as inputs to a quantified risk model spreadsheet? (approx 500 words)
Note the question excludes explanations related to project schedules or anything other than QRM. It is your understanding of the practical application that is asked here.
Practically, we have to obtain a spreadsheet ‘plug-in'. With this installed, when we click on an input cell the plug-in gives us the opportunity to select a suitable distribution, minimum, most common and maximum values. We can make these decisions for each variable in the company of every person who has something to contribute to the model. We can then run the simulation any number of times we wish to and see what the resulting output value(s) are. Because a QRM is never done with exact input data numbers we are never sure of the result, but this is a practical way of showing that uncertainty in the result.
Part 3. Specify the two risks you will use to develop your own quantified risk models.
a. The risks themselves needs to be properly defined. See what this means by reference to Module 1 and the lectures in this Module. Specific Mechanisms are not required in this. If you do include them it limits the scope of your model and you need to ask why you would want to.
b. Provide an explanation of the overall strategy you believe is required to develop each model, eg. is there bulk data available to assist you? Will you need to make a theoretical estimate of a probability in the absence of experience data? What is the unit of Exposure that is relevant? Will your risk diagram be developed from the LLC or the LWC as a starting point? I really don't know why, but many of you appeared not to have red this or seen that it was necessary to respond.
Attachment:- Quantified Risk Modelling.rar