4 Demonstration 30
4.6 Step six: Choosing the best solution 44
This paragraph is about the sixth step of the method, which is about choosing the best solution by using the Analytic
Hierarchy Process. The AHP method includes five steps: constructing a value tree, comparing attributes and
alternatives, calculating weights, calculating scores and the sensitivity analysis. The input of this step is the set of solutions that are developed in step five of the method. The output of this step should be a solution that is going to be implemented. The objectives of this step are visible in Figure 35.
First the attributes have to be determined. The attributes are extracted from the requirements and limitations to the solution. There are two types of attributes in this case, normal criteria and cut-‐off criteria. Cut-‐off criteria are so important that they cannot be skipped, but it is also hard to give scores to alternatives on cut-‐off criteria. An example of a cut-‐off criterion is that a solution has to be
conform the law; it is hard to say that alternative A is better than alternative B on that criterion, but it is also too important to skip. So cut-‐off criteria are not taken into account during the AHP, but have to be satisfied.
In Figure 36 the requirements and limitations are listed with their corresponding attribute name, whether they are a cut-‐off criterion or not and their definition. The material handling criterion is a cut-‐off criterion, because a solution can handle the material or not. The same counts for the legal criterion, something is legal or not. The requirements and limitations are relisted first.
The solution…
• has to result in a method for a precise calculation of the material and labour consumption • has to fit in a fixed budget (50.000 euro)
• cannot consume too much time of the employees when it is implemented (5 minutes) • has to handle the dimensions of the processed materials
• can be understood by all the United Springs B.V. employees • has to be supported by the United Springs B.V. employees • has to be confirm the legal and environmental legislations • can be implemented within a year
Requirement or Limitation
Attribute Name Cut-‐Off Definition
Problem solving Accuracy No Accuracy of monitoring data.
Budget Costs No Costs of alternative.
Time consumption Time No Extra time compared to current situation.
Material handling -‐ Yes -‐
Complexity Complexity No The level of complexity of new tasks.
Support Support No The level of support for the solution.
Legal -‐ Yes -‐
Implementation Implementation No The implementation complexity and time.
Figure 36. Defining the attributes for the AHP according to the requirements and limitations to the solution.
Step six
Choosing the Best Solution
Construction Value Tree
Comparison Calculating Weights Calculating Scores Sensitivity Analysis
First: The value tree
The first step of the AHP is setting up the decision hierarchy, or in other words the value tree. This value tree exists out of the attributes (criteria) and the alternative solutions. Multiple levels of attributes are possible, but in this case one level is satisfactory to keep the relations between the attributes clear. The value tree is added in Figure 37.
Figure 37. The value tree of the AHP.
Second: Make pairwise comparisons of attributes and alternatives
In this second step of the AHP the different attributes and alternatives are compared to each other. Starting with the attributes, every attribute should be compared to the other attributes to determine which of the two is: equally important (1), weakly more important (3), strongly more important (5), very strongly more important (7) or extremely more important (9). The same applies for the multiple alternative solutions. These comparisons are made by the financial controller, the research and development department and the production manager. All the comparisons are averaged to one set of comparisons; this set is attached in Appendix N.
Third: Transform the comparisons into weights and check the consistency of the decision-‐makers The third step of the AHP is transforming the comparisons into weights that sum to one (1). This is a mathematical process; all the calculations are transformed into a Microsoft Office Excel workbook so only the comparisons have to be inserted. The same counts for the consistency check. See Appendix M for the used calculation method. In Appendix O the results are attached.
Fourth: Use the weights to obtain scores for the different options and make a provisional decision The fourth step of the AHP is transforming the weights of all the attributes and alternatives into scores. These scores determine the relations between the alternatives; the alternative with the highest score is the best solution and the alternative with the lowest score is the worst solution. The winner of the AHP is the solution called M-‐Clock with a score of 0,3404. The second best solution is W-‐Clock with a score of 0,3206, the other scores are visible in Appendix O.
Fifth: Perform sensitivity analysis
The fifth and last step of the AHP is performing a sensitivity analysis. Normally this is done by a software package, but in this case this is done by use of logical sense. The average Inconsistency ratio is 0,12 and the individual scores vary between 0,0070 and 0,2475. An inconsistency ratio below 0,10 results in a choice that does not have to be doubted, unfortunately this is not the case. Also the
scores of the best alternative and the second best alternative are real close, which makes the choice even more doubtful. The four most important attributes have a better average inconsistency ratio of 0,0772.
Remarkable is that the best solution was different than the solution which was expected to be most favourable. Most people believed that weighing the material would be the best solution, not measuring the quantity of material used. This suggests that the decision makers were not too
subjective.
All in all it can be concluded that the AHP resulted in the choice of a feasible solution. The best alternative turned out to be M-‐Clock, the solution in which the material is measured and the labour consumption is determined by the use of registration terminals. The sensitivity analysis showed that the decision process was not flawless, but it is very plausible that the found solution suits United Springs B.V..