194 / 330
考量買方風險接受態度的供應商選擇：結合效用函數的簡單多屬性評比方法
194
Infrastructure includes management, financial stability, and reputation of the supplier.
Infrastructure management involves suppliers’ commitment to improve as well as its ability
and performances. Moreover, suppliers should have long-range planning and contingency
plans in case of disruptions. Financial stability considers if the suppliers have good credit
and if they are financially healthy (they can meet their debt and financial obligations).
Financial stability can be quantified by evaluating suppliers’ liquidity ratios, turnovers and
interest coverage. Unstable suppliers create a risk of possible disruptions in the long run
which cause problems in the supply flow (Coyle et al., 2013). Attitude refers to the
relationship between the supplier and its customers (Coyle et al., 2013). This can be
quantified in the years of working together and comfort level (Barla, 2003).
3.2 SMART
SMART is one of the commonly used methods in multi-attribute decision-making
because of its simplicity. Edwards (1977) and Olson and Wu (2010) described the steps of
SMART as follows. First, one should identify the decision-maker and determine the possible
alternatives. Then, the attributes and their criteria of interest to the decision-maker are
defined. Finally, the dimensions or attributes are ranked according to their (relative)
importance. The importance of attributes can be represented by their weights. In this
research, two types of weights are used: experts’ weight and Rank Order Centroid (ROC)
weight. In ROC, the dimensions are ranked in order of importance before giving them
weights. Equation (1) shows the formula for calculating Rank Order Centroid (ROC) weights
proposed by Barron and Barrett (1996).
(1)
where
W
i
is the weight of criteria/attribute
i
, and
K
is the number of criteria. According
to their empirical results, the ROC weights are close to experts’ weights. Thus, when experts’
weights are not available, the ROC weights may be used as substitutes. After getting the
weights, the next step is to sum and normalize the weights. Then, each alternative is
measured according to the given attributes/criteria. When all attributes are measured, the
utility of each alternative is calculated as follows:
(2)