SMART Flashcards
what are the problems with multi-attribute decision making?
attribute selection, comparing alternatives on attributes, and combining performance data to make the selection
what are the criteria for a value tree?
completeness, operationality, decomposability, absence of redundancy, minimum size.
what are the three types of attribute SMART defines?
monetary, those with a natural numeric scale, those without a natural numeric scale
how does SMART measure attributes with a natural numeric scale?
value function
how does a value function work?
worst choice gets 0, best gets 100. place others inbetween proportionally
how does SMART measure attributes with no numeric scale?
direct rating
how does direct rating work?
decision maker’s best and worst choices are given 100 and 0. Then ask them to place the rest on the scale 0-100.
why are scales of 0-100 used in SMART when comparing alternatives on attributes?
so every attribute’s contribution to the overall choice is equally weighted, and swing weights can be applied to an equal playing field.
how do swing weights work?
note the actual values of the best and worst choices for each attribute. then order them by asking ‘if you have the worst of everything, which one would you choose to change to the best?’. keep going until you’ve done them all, best attribute gets 100 and worst gets 0. use direct rating for the rest. the total can be more than 100.
how do you work out the total benefit points for an option with SMART?
sum of (attribute score * normalised attribute weight).
how do you create an efficient frontier using SMART cost-benefit options?
plot the costs and benefits against each other, the efficient frontier is the options that aren’t dominated by any other.
how do you work out cost of a value point in SMART?
identify an attribute the decision maker can say they’d pay £x to swing from best to worst. then do £x/(normalised weight*100) to find how much they’d pay per point.
what are the advantages of SMART?
it breaks the problem down and makes it easier to handle, allows decision maker to address all the information and objectives, can carry out what if analysis, helps with information gathering, provides defensible rationale, challenges intuition, helps consistency and it’s quick/transparent compared to other methods.
what are the disadvantages of SMART?
it’s time-consuming, the ratings and weights can depend on the question framing, there is an assumption of linearity in the value functions and when working out £/value point.
what is the difference between decomposability and absence of redundancy?
decomposability asks to ensure attributes are independent from each other, absence of redundancy just ensure there is no duplication.
