CLASIFICACIÓN DE LOS SUSTRATOS

In the preparation and engagement phase, the determined influences of the decision attributes

were used to design the benefits and costs network structures. Figure 6.1 shows the benefits

control network model designed for the decision-making problem in this Chapter. The costs

control network is not shown here. However the interactions in both networks were used to

Pertinently, the revised goal-rating document used in eliciting stakeholder judgements with the Analytic Hierarchy Process (AHP) (Section 0) also contained the Analytic Network Process (ANP) judgement matrices with respect to dependencies/interdependencies of decision attributes. The new contents of the revised goal rating document were considered necessary to enable a single engagement of each participant-stakeholder. This manages the difficulty of expending more time and money in re-interviewing the stakeholders to implement the ANP, given stakeholders’ busy work schedule and different location in New Zealand. Other contents of the goal rating document have been listed in Section 4.3.1.1, and a sample of the document is included in Appendix 2(d). The stakeholder views on the decision cluster/elements/options and their associated dependences were extracted at the same response and rating phase of the stakeholder engagement process explained in Section 0. This entailed that the AHP and ANP type of paired comparison judgements were extracted at the same time from each participant-stakeholder.

Figure 6.1. Structural fire design decision 'benefits' criteria network model.

The loops on the safety, environmental and societal clusters in Figure 6.1 indicate that inner

dependencies exist in the clusters, such as observed with C2, C4 and C5 in Figure 3.3. The one

and two-way arrows among the clusters in Figure 6.1 indicate outer dependencies same as C1

– C5 in Figure 3.3.

6.3.2.1. Determination of the cluster weights (cluster matrix)

The cluster weights of the benefits and costs network models were also determined at this

point based on relative influences/interactions of each cluster in the networks using the

Safety cluster FRA, SFR, PF1, CDD, FFO, FSC, MA Environmental cluster ES, EAC

Fire protection options cluster ITC, BST, CES,

SCM, UPS

Societal cluster

BA, HC, ASI, BRA, BUF, HS, PF2

pairwise comparison and AHP-weighting procedure. Table 6.1(a) shows the weights of the

clusters in the benefits control criteria network. From the benefits network model (Figure

6.1), the elements in the options cluster were considered as not inner dependent. This consideration was depicted with the zero influence weight for the options cluster in the matrix

(Table 6.1{a}); zero was also entered for environmental as it had no outer influence on the

safety and societal clusters. The costs network cluster weights are shown in Table 6.1(b), here the interaction is between the economy cluster and the fire protection options’ cluster which follows the same explanation given above for Table 6.1(a).

Table 6.1. Cluster matrixes (a) benefits control network (b) costs control network

(a)

Safety Environmental Societal Fire protection options

Safety 0.42 0.00 0.06 0.33

Environmental 0.06 0.50 0.06 0.33

Societal 0.11 0.00 0.44 0.33

Fire protection options 0.42 0.50 0.44 0.00

(b)

Economy Fire protection options

Economy 0.25 1.00

Fire protection options 0.75 0.00

6.3.2.2.Stakeholder engagement and paired comparison ratings

This research considered the response and rating phase of the stakeholder engagement process as a continuous phase given that the decision analyses stages and tools (GMM-AHP and W/GMM-ANP) had been automated. Therefore, at this research stage, 42 chartered and experienced stakeholders within the building and fire industry in New Zealand have been interviewed. This includes the previous sample set (i.e. 36 stakeholders) in Chapter 5. As mentioned earlier, all participant-stakeholders in this research are interviewed once using the same goal rating document whereby their views are elicited given the AHP and ANP type of questions, decision models and judgement matrices. The fire design stakeholder categories and rate of individual participation are shown in Figure 6.2.

Although it would be desirable to increase the number of participants in this study, the 42 participant-stakeholders from the 12 fire design stakeholder categories are considered sufficient to carry out the stakeholder decision analysis. Also, the sample-set is sufficient to demonstrate the viability of the W/GMM-ANP in comparison with previous attempts to apply AHP in fire safety studies discussed earlier in Section 3.1.1.2. There is no fixed amount of stakeholder/s or decision-maker/s’ judgements needed to achieve the best-ranking result through AHP or ANP. As discussed in Section 3.2.1, the techniques were initially developed as single decision-maker techniques before the advent of W/GMM and W/AMM for group decision aggregation. Therefore, in a general case study as carried out here, the number of stakeholders involved may be considered as the more, the better.

Figure 6.2. The fire design stakeholder categories and individual participation.

Based on the fundamental scale (Saaty’s reciprocal scale) and ANP question in the goal- rating document, each stakeholder carried out the pairwise comparison rating of the decision elements at the response and rating phase of the stakeholder engagement. Table 6.2 shows the

judgement matrices of the outer dependence of clarity in design detail and specifications

(CDD) on all stakeholder involvement in design (ASI), building regulation approval (BRA)

and health and safety (HS) (described in Section 6.3.1.1) from the two stakeholders in building owners’ category (BDO1 and BDO2).

2 3 3 1 2 4 3 7 9 1 3 4 Stakeholder categories

Table 6.2. Building owners' judgements of some social cluster elements influence on 'clarity in design details and specifications'.

(a) Judgement matrix from BDO1 (b) Judgement matrix from BDO2.

CDD ASI BRA HS CDD ASI BRA HS

ASI 1 1/9 1/9 ASI 1 1/5 1/5

BRA 9 1 1 BRA 5 1 1

HS 9 1 1 HS 5 1 1

Here, building owner 1 (BDO1) and building owner 2 (BDO2) have the same view that

building regulation approval (BRA) and health and safety (HS) have equal influence on

clarity in design details (CDD) as they rated both elements as ‘1’ in Table 6.2 (a) and (b).

However, their judgements differ regarding the influence of all stakeholder involvement in

design (ASI) on CDD. BDO1 [Table 6.2 (a)], rated BRA and HS as having more influence on CDD than ASI with the value ‘9’ and its reciprocal value, 1/9 which means that ASI has less influence on CDD relative to BRA and HS. On the other hand, BDO2 judged BRA and HS as having much more influence on CDD with the value ‘5’ and vice versa. As previously seen in

Table 5.2 (i.e. structural engineers’ judgements), Table 6.2 also shows the different views of

fire design stakeholders within the same category. Similar variations were also observed in other stakeholder categories not shown here.

During the stakeholder interviews, the building owners commented that they cared more about achieving building consent. Hence, they would naturally insist that the structural and fire engineers ensure clarity in their designs and also to reduce uncertainties during the construction of the steel-framed building. Importantly, the building owners informed the researcher that they do not prefer all the stakeholder categories in this research to participate in design decision-making processes especially environmental professionals and end-users. Their views were based on managing potential conflicts in a decision-making process. These views were reflected in their judgements, as shown in Table 6.2. Other stakeholder views on some decision elements and competing options had been summarised earlier in Section 5.2.2.1.

The determination of the cluster weights and the collection of individual judgements of the 42