The model formulation process includes four features that help derive a model:
I. Decision,
II. Outcome,
III. Structure
IV. Data
The model formulation process is a framework of the type’s variables (decisions) as input to a model and the possible outcomes derived from those decisions. The formulation process also defines structure. Structure is the logic and math that ties the various parts of a model together. Lastly there is data; data can be observed information (AKA raw or empirical) such as actual counted items produced in production or it can be mathematical assumptions - this information that is input to our model.
Here is an example:
Decision:
Increase production of widgets in anticipation of expected sales.
Outcome:
Production is increased and the amount widgets on store shelves are doubled. A possible outcome can be a shortfall of sales and lower than projected income from sales. Costs associated with increased manufacturing production are not recouped. Unanticipated loss of popularity of widgets could be one possible cause. Another possible outcome is widget popularity has increased and the production estimates have erred on the low side. The net effect is that widgets have flown off the shelves and public demand has stores clamoring for more widgets. While all costs associated with increased widget production are recouped and profit expectations where met; company analysts misjudged market conditions and a window to satisfy public demand is missed as unhappy would be widget owners lose interest. However; all is not lost as there a significant widgets on back-order.
Structure:
These could be simple mathematical production formulas with input variables such cost of labor, how many widgets per hour produced, cost of materials, facility overhead etc..
Data:
I. Raw/Empirical
a. Widget owner opinion survey results.
b. Public opinion surveys regarding widgets
II. Estimates
a. Projected sales
Sunday, July 11, 2010
Summarize section 2.3.2 including the influence chart
Summarize section 2.3.2 including the influence chart. This section takes the influence chart from the previous section (“a pricing decision)” and adapts it to fit “an income statement as an influence chart” This is achieved by further decomposing the variable called “variable cost” to a variable called “quantity sold” The variable called “total revenue” is also decomposed into “quantity sold”. Total revenue has the symbol for decision added as “total revenue” is decomposed to a decision called “price”. “Quantity sold” is also decomposed to “price” and additionally the fixed input parameter called “elasticity”. The section explains that the chart represented in the section is good for a static balance sheet as representative of the historical performance of the example company because all the variables are predetermined. It further explains that it would not suffice if the goal was to attempt projections and estimates into the future. The reasons given explain that figure 2.4 is a static framework and not a model for the future. To convert a static statement into a model; variables such as Quantity Sold need to evolve over time.
Summarize section 2.3.1 including the influence chart.
A. Section 2.3.1 Summary including the influence chart. This section walks through the decomposition process in building an influence chart - a pricing decision is used as the example for this section. It demonstrates that variables can be decomposed into sub-variables and fixed input parameters some sub-variables with their own set of variables and fixed input parameters. It begins with an objective “retained earnings” and decomposes it into some variables known to be needed in obtaining that objective, in this case “profit after taxes”, “profit before taxes” and “expenses”. The section outlines guidelines for constructing a chart:
B. Decide which variable this is; ask what single variable the decision maker will use to measure the success of a plan of action.
C. Decompose the outcome to measure into a small set of variables that determine it directly. Each of these variables should be independent of the others, and together they should be sufficient to determine the result.
D. Take each variable in turn and repeat this process of decomposition. For each variable, ask, “What do I need to know to determine…?”
E. Identify input data and decisions as they arise
F. Make sure that each variable appears only once in a diagram.
G. Highlight special types of elements with consistent symbols. In this section the following symbols were used:
a. Symbol for objective
i. An octagon is used.
b. Symbol for variable
i. A landscape oval is used
c. Symbol for fixed input parameter.
i. Triangles are used
B. Decide which variable this is; ask what single variable the decision maker will use to measure the success of a plan of action.
C. Decompose the outcome to measure into a small set of variables that determine it directly. Each of these variables should be independent of the others, and together they should be sufficient to determine the result.
D. Take each variable in turn and repeat this process of decomposition. For each variable, ask, “What do I need to know to determine…?”
E. Identify input data and decisions as they arise
F. Make sure that each variable appears only once in a diagram.
G. Highlight special types of elements with consistent symbols. In this section the following symbols were used:
a. Symbol for objective
i. An octagon is used.
b. Symbol for variable
i. A landscape oval is used
c. Symbol for fixed input parameter.
i. Triangles are used
The steps to derive a problem from mess.
The steps to derive a problem from mess:
A mess is as the name implies a jumble of symptoms, data, pressures, opportunities etc – a mess is not clearly defined and may contain multiple as yet unidentified problems. A problem is a well defined situation capable of resolution. To derive and identify a problem from a mess one must employ critical thinking skills and remove any bias and emotion from the process. There may be several problems nested within a mess that may require identification and subsequent resolution. It may be the case that only the most serious problems will be addressed; this could be due to several factors such as time to resolution, cost to implement the solution (cost in money and cost in human capitol) and public perception. In the world of politics and government; public perception can take a good solution and table it in favor of a less preferred but better received option.
Problems can be well structured with the following characteristics:
i. The objectives are clear
ii. Any needed assumptions are obvious
iii. All needed data are available
iv. The logical structure behind the analysis is well understood.
By contrast an ill structured problem has all of the opposite characteristics as those listed above. As mentioned in the six step process; divergent and convergent thinking can be employed in identifying and solving the problem(s). Divergent though is a brainstorming where no idea or concept is off the table. Convergent thinking takes those ideas and pares down that list into more focused areas purposed with finding one best solution.
A mess is as the name implies a jumble of symptoms, data, pressures, opportunities etc – a mess is not clearly defined and may contain multiple as yet unidentified problems. A problem is a well defined situation capable of resolution. To derive and identify a problem from a mess one must employ critical thinking skills and remove any bias and emotion from the process. There may be several problems nested within a mess that may require identification and subsequent resolution. It may be the case that only the most serious problems will be addressed; this could be due to several factors such as time to resolution, cost to implement the solution (cost in money and cost in human capitol) and public perception. In the world of politics and government; public perception can take a good solution and table it in favor of a less preferred but better received option.
Problems can be well structured with the following characteristics:
i. The objectives are clear
ii. Any needed assumptions are obvious
iii. All needed data are available
iv. The logical structure behind the analysis is well understood.
By contrast an ill structured problem has all of the opposite characteristics as those listed above. As mentioned in the six step process; divergent and convergent thinking can be employed in identifying and solving the problem(s). Divergent though is a brainstorming where no idea or concept is off the table. Convergent thinking takes those ideas and pares down that list into more focused areas purposed with finding one best solution.
The principles/steps/symbols in building influence charts
The principles/steps/symbols in building influence charts :
An influence chart is something of a free form visual aid that helps us conceptualize a model. Similar to a flow chart but not as step by step – does not delve into the minutia of each element of a step by step process – if A than B; if B is true go to C if not repeat A et… A flow chart can be based on Boolean logic. An influence chart is a high level thought flow, about areas of concern, which should be given treatment in some way for a given outcome to be achieved. Each outcome leads to another until some final result is achieved. While the outcome of a flowchart is known; it is constructed in order from beginning to end. In contrast; an influence chart starts at the end (objective) and works outward. This is achieved by decomposing the objective into variables needed to obtain the objective and then further decomposing each variable as needed to arrive at the component variables for the parent variable. Each variable might have fixed input parameters and random input parameters. As an example the variable “profit” might have a sub-variable of “expenses”; “expenses” might have fixed input parameters such as “cost of goods”, “depreciation”, and “interest”. A variable could also be some function of the manufacturing process with sub variables such as temperature and humidity. The symbols used are up to the creator or organization - the important factor is that they are consistent thorough out the chart.
An influence chart is something of a free form visual aid that helps us conceptualize a model. Similar to a flow chart but not as step by step – does not delve into the minutia of each element of a step by step process – if A than B; if B is true go to C if not repeat A et… A flow chart can be based on Boolean logic. An influence chart is a high level thought flow, about areas of concern, which should be given treatment in some way for a given outcome to be achieved. Each outcome leads to another until some final result is achieved. While the outcome of a flowchart is known; it is constructed in order from beginning to end. In contrast; an influence chart starts at the end (objective) and works outward. This is achieved by decomposing the objective into variables needed to obtain the objective and then further decomposing each variable as needed to arrive at the component variables for the parent variable. Each variable might have fixed input parameters and random input parameters. As an example the variable “profit” might have a sub-variable of “expenses”; “expenses” might have fixed input parameters such as “cost of goods”, “depreciation”, and “interest”. A variable could also be some function of the manufacturing process with sub variables such as temperature and humidity. The symbols used are up to the creator or organization - the important factor is that they are consistent thorough out the chart.
The six stage problem solving process.
The six stage problem solving process is a process that begins with a “mess” and ends with the implementation of a solution. The six step process is achieved by one of two styles of thinking:
Divergent
Convergent
Divergent thinking is akin to brainstorming. It allows for many ideas, directions and solutions to be formulated. Convergent thinking zero’s in on the best of the divergent thinking and begins to look deeper. The six stages are:
Exploring the mess
What problems or opportunities exist?
Are there stated or unstated goals?
Produce a description of the situation.
Identifying the key facts and data.
Searching for information.
Information is gathered from many sources.
What are symptoms and causes?
Information is culled and pared down to what is relevant to the situation.
What measures seam appropriate?
What remediation is available?
This phase can be considered completed once we have found and organized relevant information and have made some initial hypothesis about root cause and potential solutions.
Identifying a problem
Work towards a single problem statement or multiple sub-problems.
Which is the most important problem in this situation?
Is this problem like others we have dealt with?
What are the consequences of a broad versus narrow statement?
This stage is complete when a working problem statement has been produced.
Searching for solutions.
As the name implies; this stage involves finding potential solutions. The goal is a short list of candidate solutions. To get there we could consider the following:
What decisions are open to us?
What solutions have been implemented in similar situations?
How is the candidate solutions linked to outcomes of interest?
Evaluating solutions.
This stage is the bottom of the funnel – the point at which the preceding stages converge. The goal of this stage is to produce a working solution with supporting justification.
Implementing Solutions
The goal of this stage is to produce an implementation plan and execute enough of it to evaluate its impact and chances for success. It is worth noting that the success is as much political as technical. The current crude oil leak in the Gulf of Mexico is a potent reminder of the public relations and political agendas that could impact this and any of the other 5 stages.
Is the solution right for the environment?
Is it right for BP?
Is it right for the local economy?
Is it right for the government?
Subscribe to:
Posts (Atom)