**5. Operation of the prototype system**

The developed prototype system provides a decision support tool to decision-maker by aiding in the selection of the most qualified contractor to upgrade internet infrastructure. The system allows for input from users in the decision-making process. The prototype system:

**•** provides a clear and structured framework of implementation of the decision-making process, which assists the user in the selection of the most qualified contractor to upgrade the internet infrastructure based on the criteria and sub-criteria included in the AHP model



#### **5.1. System requirement**

development of the National Centre for Technology Management (NACETEM), an agency of the Federal Ministry of Science and Technology in Nigeria, is used as illustration. NACE‐ TEM is an agency vested with the mandate of training and developing middle- to high-level manpower and conducting policy research in the areas of science, technology, and innova‐

To select the most suitable contractor for the infrastructural development of the agency base on AHP methodology, information was gathered from all the departments in the organiza‐ tion on the selection process and later carefully analyzed so that standard criteria and subcriteria could be established and later adjusted with respect to the general goal. Once the problem has been defined as selecting the best contractor, the process starts with the design of the prototype system, which includes designing the system architecture and identifying the implementation and operational framework. The knowledge acquired through the knowl‐

Judgment of the results by the expert (i.e., tenders board) and evaluation of the prototype are carried out, whereby the effectiveness of the software and hardware is checked. If the results or findings from the evaluation require some improvement, the prototype is modified and

The aim of the evaluation is to determine the usability and functionality of the finished prototype. This would be achieved through interviews and questionnaire administered to the

The developed prototype system provides a decision support tool to decision-maker by aiding in the selection of the most qualified contractor to upgrade internet infrastructure. The system

**•** provides a clear and structured framework of implementation of the decision-making process, which assists the user in the selection of the most qualified contractor to upgrade the internet infrastructure based on the criteria and sub-criteria included in the AHP model

allows for input from users in the decision-making process. The prototype system:

tion management for all tiers of government and the private sector.

300 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

edge acquisition process is represented in the prototype in four main steps:

**1.** Developing the hierarchy as shown in **Figure 1**

**2.** Pairwise comparison of criteria and sub-criteria

**4. Evaluation of the prototype system**

**5. Operation of the prototype system**

**4.** Measuring inconsistency in decision-maker's judgments

**3.** Synthesis of the AHP model

redesigned as appropriate.

experts (tenders board).

The prototype system operates on personal computer running window XP or later version for smooth operation. However, it can still work with lower versions of window XP. It requires Expert Choice 11 which is the current version, Microsoft Office 2003 orlater version (MS Word, MS Excel, MS Access, etc.) to be installed. It is also advisable to have relatively large memo‐ ry and storage capacity (say, 1 G of RAM and 100 G of hard disk size).


**Figure 2.** Expert Choice welcome window view panel that is divided into three major panes.

#### *5.1.1. Starting the prototype system*

The prototype system is stored as an Expert Choice file called "contractor selection.ahpz" and held in directory with specific name, e.g., "C:\users\Ola-dotun\document\contractor selection.ahpz."

To start the application, click on Start button on the desktop, then select All Programs, and click on Expert Choice or double-click on Expert Choice from the desktop to open the application (**Figure 2**); if one is designing a fresh model, then click on Create New Model, select Structuring, and click Ok. The application will display a dialog box where you can type the name of the model and save to Start Building a New Model.



**Figure 3.** Contractor selection system using Expert Choice.

#### *5.1.2. Assigning judgment in pairwise comparison*

One of the main strengths of AHP is the use of pairwise comparisons to derive accurate ratio scale priorities. The pairwise comparison process compares the relative importance, prefer‐

ence, orlikelihood of two elements with respect to each other. A judgment is made as to which is more important and by how much. Judgment about the relative importance of criteria is made with respect to the parent node in the hierarchy (either the goal or a higher-level criterion). Judgments about the relative preference of alternatives are made with respect to each criterion. You can make judgments/pairwise comparisons starting with the goal and working down to the alternatives (top-down), or you can make judgments about the alterna‐ tives before making judgments about the objectives (bottom-up). The bottom-up approach is usually better because the insights you gain about the trade-offs among the alternatives will help in making judgments about the importance of the objectives [8]. Expert Choice pro‐ vides judgment in three different forms, namely pairwise numerical comparison, pairwise verbal comparison, and pairwise graphical comparison (**Table 1**).

*5.1.1. Starting the prototype system*

selection.ahpz."

the system.

The prototype system is stored as an Expert Choice file called "contractor selection.ahpz" and held in directory with specific name, e.g., "C:\users\Ola-dotun\document\contractor

To start the application, click on Start button on the desktop, then select All Programs, and click on Expert Choice or double-click on Expert Choice from the desktop to open the application (**Figure 2**); if one is designing a fresh model, then click on Create New Model, select Structuring, and click Ok. The application will display a dialog box where you can type the

**1.** *The Tree View pane*. The hierarchy displayed in this (contractor selection model) consists of five main criteria and fifteen sub-criteria with the goal being to select the best contrac‐

**2.** *Alternative pane*. The list of alternative is displayed here. That is, the four companies that

**3.** *Information document pane*. This includes information on operating the system and links to other information document files, which may be of importance for easy understanding of

One of the main strengths of AHP is the use of pairwise comparisons to derive accurate ratio scale priorities. The pairwise comparison process compares the relative importance, prefer‐

name of the model and save to Start Building a New Model.

is being considered in handling the project.

**Figure 3.** Contractor selection system using Expert Choice.

*5.1.2. Assigning judgment in pairwise comparison*

tor for the upgrading of internet infrastructure (**Figure 3**).

302 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions


**Table 1.** Rating of pairwise comparison of a pair of criteria and sub-criteria using AHP preference scale.

For instance, the user makes judgments about the preference of a contractor (company) with respect to the criterion, "average annual turnover." The steps to take include:


The lower cell is called the active cell and contains the numerical (in black) representa‐ tions by which the elements (i.e., companies) displayed along the rows are more prefer‐ red to those displayed along the columns in a matrix format, based on the comparison of average annual turnover criterion, while any judgment shown in red indicates that the column element is preferred to the row element.

**4.** The above process (step 3) is repeated until all the companies have been compared, by pairing them in twos, with respect to their average annual turnover.

Inconsistency: Possible errors and actual inconsistency in judgment are identified by the inconsistency measure shown in the bottom-left cell of the matrix. For the decisionmaker's judgment to be reasonably consistent, the inconsistency ratio must be less than 0.1. The decision-maker should only change an inconsistent judgment if he truly feels that the initial comparison was in error.

**5.** Recording judgments and calculating priorities.

When all judgments have been made, one is prompted to record judgments and carry out necessary calculation. Select YES so as to return to the Model View.

The priorities of alternatives with respect to average annual turnover are automatically computed and displayed in the Pane of the Model View. **Figure 5** shows the priorities of the companies with respect to "average annual turnover."

In a situation where the resulting relative priorities do not reflect the decision-maker's feelings, the pairwise comparison process is repeated. These can be achieved either by using the verbal mode or by switching over to either the graphical mode or numerical mode. To obtain the rightresult, the user should click on the tab of one ofthe three pairwise comparison tabs. The user then selects the mode that he feels will make the most sense and later uses the selected mode to obtain the judgment for the current set of criteria (or sub-criteria) being compared.


To assign judgment to the sub-criteria against criteria as well as to criteria against the goal, the user (decision-maker) must change the comparison type from "preference" to "importance."

To change the assessment type:



**Figure 4.** The verbal comparison window.

The lower cell is called the active cell and contains the numerical (in black) representa‐ tions by which the elements (i.e., companies) displayed along the rows are more prefer‐ red to those displayed along the columns in a matrix format, based on the comparison of average annual turnover criterion, while any judgment shown in red indicates that the

**4.** The above process (step 3) is repeated until all the companies have been compared, by

Inconsistency: Possible errors and actual inconsistency in judgment are identified by the inconsistency measure shown in the bottom-left cell of the matrix. For the decisionmaker's judgment to be reasonably consistent, the inconsistency ratio must be less than 0.1. The decision-maker should only change an inconsistent judgment if he truly feels that

When all judgments have been made, one is prompted to record judgments and carry out

The priorities of alternatives with respect to average annual turnover are automatically computed and displayed in the Pane of the Model View. **Figure 5** shows the priorities of

In a situation where the resulting relative priorities do not reflect the decision-maker's feelings, the pairwise comparison process is repeated. These can be achieved either by using the verbal mode or by switching over to either the graphical mode or numerical mode. To obtain the rightresult, the user should click on the tab of one ofthe three pairwise comparison tabs. The user then selects the mode that he feels will make the most sense and later uses the selected mode to obtain the judgment for the current set of criteria (or

**6.** Processes 1–5 are repeated until all comparisons for sub-criteria (15 nodes) and criteria (5

To assign judgment to the sub-criteria against criteria as well as to criteria against the goal, the user (decision-maker) must change the comparison type from "preference" to

pairing them in twos, with respect to their average annual turnover.

304 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

necessary calculation. Select YES so as to return to the Model View.

the companies with respect to "average annual turnover."

column element is preferred to the row element.

the initial comparison was in error.

sub-criteria) being compared.

To change the assessment type:

nodes) have been made.

**7.** Assignment of judgment.

**•** Select assessment.

**•** Select importance.

"importance."

**•** Select type.

**5.** Recording judgments and calculating priorities.


**Figure 5.** Derived priorities of the alternatives with respect to average annual turnover.

#### *5.1.3. Synthesizing and measuring inconsistency*

Tree View in both graphic and numerical form.

To examine the synthesis, select Synthesis, with respect to the goal to produce the display shown in **Figure 6**, while **Figure 7** shows the Synthesis window displaying the charts.

**Figure 6.** Model View showing the synthesized results with respect to the goal.

**Figure 7.** Synthesis window.

In Expert Choice, the "Distributive Mode" and "Ideal Mode" are two synthesis methods that can be used to derive the results. Distributive Synthesis is used when the users are interest‐ ed in prioritizing alternatives from which they may pick more than one alternative, while the ideal Synthesis should be used when one is interested in only one alternative, and the remaining alternatives are no longerrelevant. The difference in results obtained using the ideal or distributive synthesis modes is usually negligible and more of theoretical than practical interest.

#### *5.1.4. Sensitivity analysis*

*5.1.3. Synthesizing and measuring inconsistency*

Tree View in both graphic and numerical form.

306 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

**Figure 6.** Model View showing the synthesized results with respect to the goal.

**Figure 7.** Synthesis window.

To examine the synthesis, select Synthesis, with respect to the goal to produce the display shown in **Figure 6**, while **Figure 7** shows the Synthesis window displaying the charts.

> Expert Choice provides tools for performing sensitivity analysis. Sensitivity analysis dis‐ plays how the alternatives change priorities if there is a change in the importance of the criteria or sub-criteria graphically. Starting with the goal node, sensitivity analysis displays the sensitivity of the alternatives with respect to all the criteria below the goal. Since the AHP model in use as illustration has more than three levels, sensitivity analysis can also be performed from the nodes below the goal to show the sensitivity of the alternatives with respect to the criterion and sub-criterion [8]; and according to Ref. [9], when performing a sensitivity analysis, the user may change the priorities of the criteria and observe how the priorities of the alternatives would change.

**Figure 8.** Graph of sensitivity analysis.

There are five types of sensitivity analysis that can be carried out in Expert Choice, namely, performance, dynamic, gradient, head-to-head, and two-dimensional plot. The first three are often used to perform the sensitivity analysis because you can dynamically vary them by dragging the objective bars.

To perform sensitivity analysis from Tree View, click on the goal or criteria or sub-criteria that you want to perform analysis on with respect to the alternatives, and select sensitivity graph, then click the type of sensitivity analysis you want to perform.


**Figure 9.** Graph of dynamic sensitivity analysis.


**Figure 10.** Graph of gradient sensitivity analysis.
