Building a Hierarchy Model for Evaluating Different MAN Access Strategies

The emerging broadband service demands in the MAN have forced telecom carriers to evaluate different access strategies in order to meet customer needs. Many business criteria need to be considered while selecting an access strategy. However, many business criteria are fuzzy in nature and in mutual conflict. For example, if an access strategy only focuses on the equipment cost, the issue of equipment retrofit and backward compatibility could have a negative impact on the telecom carrier network operation and maintenance costs. Therefore, how to select a business strategy that can reach a compromise solution becomes a critical issue.

4.2.1 Building a hierarchy strategy model

The PATTERN (Planning Assistance through Technical Evaluation of Relevance Number) method and concept (NASA PATTERN, 1965, 1996; Tzeng, 1977; Tzeng and Shiau, 1987;

Tzeng. et al., 1992; Tzeng and Teng, 1994; Tang, et al., 1999) to build a hierarchy strategy system for evaluating business strategies was employed in this paper. The PATTERN procedures include three steps: (1) scenario writing, (2) building a relevance tree, and (3) evaluation. In this section, we focus on scenario writing and building a relevance tree. The business issues can be classified into three categories: (1) Network Services (2) Network Equipment Costs, and (3) Network Operations. Based on the literature review and experience, relevance trees are used to create hierarchy strategies to identify the critical business issues

and criteria using scenario writing. The elements of relevance trees become a relevance set consisting of statements derived from “goal” through aspect, objective, planning to implementation. This kind of system corresponds with evaluation processes for telecom carriers to evaluate critical network access strategies/issues (as shown in Figure 3).

Goal Aspects Criteria/Objectives Strategies

Figure 3. Relevance System of Hierarchy Strategies for Access Technology

4.2.2 Access strategy for MAN

Based on different focuses and criteria, different access strategies can be selected for consideration. However, since the criteria for each network access strategy are in conflict with each other, it is important for telecom carriers to consider all aspects of the business environment so that carriers can balance their short-term and long-term goals. Two different strategies are summarized in Table 9.

Compatible with existing

Table 9. MAN access strategies for the telecom carriers

Criteria/Issues Strategies

• Telecom carriers need to provide proven carriergrade scalable broadband solutions and service for customers.

• It is critical to maintain network operations through comprehensive network management information.

• Leveraging existing SONET/SDH infrastructure to generate revenue is important, particularly during the telecom downturn.

A. Focus on SONET/SDH and ATM access technology:

Telecom carriers continue to enhance and improve existing SONET/SDH and ATM networks and their respective operations support systems to offer quality broadband services.

• Telecom carriers should offer bandwidth flexibility to customers.

• The convergence of voice and data service will be the major business drivers for future broadband services

• Internet IP protocol and WDM optical bandwidth will continue to stimulate the Ethernet deployment and to continue to drive down the cost of Ethernet technology.

B. Focus on Gigabit Ethernet access technology:

To offer broadband service, telecom carriers will aggressively deploy Gigabit Ethernet in the MAN to take advantage of potential benefits offered by Gigabit Ethernet.

4.2.3 Fuzzy MCDM Method

In a simple environment or using a single measurement index, the traditional minimum cost, maximum profit or cost efficiency methods can be employed to conduct alternate evaluations.

However, in an increasingly complex and diversified decision making environment, there is much correlated information that needs to be analyzed and traditional analysis is not suitable for problem solving (Tzeng, et al., 1992; Tzeng & Tsaur, 1993; Tzeng and Teng, 1994; Tsaur et al., 1997; Tang et al., 1999). Therefore, this research uses the MCDM to evaluate different MAN access strategies.

Since evaluators may have different perceptions on different objectives and criteria, in terms of their importance and possible adverse consequences, evaluation is conducted in an uncertain and fuzzy environment. This fuzzy evaluation design allows evaluators to express their opinions in a fuzzy expression manner. For these reasons, the Fuzzy MCDM was selected to conduct this evaluation.

4.2.4 The process of evaluating the hierarchy strategies The evaluation process includes two steps.

4.2.4.1 Evaluating the weights for the hierarchy relevance system using AHP (Analytic Hierarchy Process)

The AHP weighting is determined by evaluators who conduct pairwise comparisons. This matches the criteria to discover the comparative importance of each. If there are evaluation criteria/objectives, then the decision-makers have to conduct a pairwise comparison.

Moreover, the relative importance derived from these pairwise comparisons allows a certain degree of inconsistency within a domain. Saaty used the principle eigenvector of the pairwise comparison matrix derived from the scaling ration to find the comparative weight among the criteria of the hierarchy system for the enterprise business strategies.

4.2.4.2 Obtaining performance values

The evaluators choose a score for each business strategy based on their subjective judgment.

In this way, we can use the methods of fuzzy theory to estimate the achievement level of each strategy in a fuzzy environment. Since Zadeh introduced fuzzy set theory (Zadeh, 1965), and Bellman and Zadeh (1970) described the decision-making method in fuzzy environments, an increasing number of studies have dealt with uncertain fuzzy problems by applying fuzzy set theory. The application of fuzzy theory to obtain performance values can be described as follows:

1. Fuzzy Set: Fuzzy numbers are a fuzzy subset of real numbers, and they represent an expansion of the idea of a confidence interval.

2. Linguistic Variable: According to Zadeh (1975), it is very difficult for conventional quantification to reasonably express complex and/or hard-to-define situations; thus, the notion of a linguistic variable is necessary in such situations. A linguistic variable is a variable whose value are words or sentences in a natural or artificial language. For

example, the expression “maximize network bandwidth and QoS ” or “reduce operations cost” represents a linguistic variable in the context of this study. Linguistic variables may take on effect-values such as “very high (very good), “high (good)”, “fair”, “low (bad)”, and “very low (very bad)”. The use of linguistic variables is rather widespread at present and the linguistic effect values of enterprise business strategies found in this study are primarily used to assess the linguistic ratings given by the evaluators. Furthermore, linguistic variables are used as a way to measure the achievement of the performance value for each criteria/objectives (Fig. 4).

Figure 4. The membership function of the five levels of linguistic variables (hypothetical example)

3. Fuzzy Multiple Criteria Decision-Making: Bellman and Zadeh (1970) were the first to probe into the decision-making problem under a fuzzy environment, and they heralded the advent of Fuzzy MCDM. This study uses this method to evaluate various business strategies and ranks each strategy according to its score.

4. Fuzzy synthetic decision: The weights of each criteria/objective of the MAN access strategies as well as fuzzy performance values have to be integrated by the calculation of fuzzy numbers so they can be assigned a fuzzy performance value for the integral evaluation. This procedure is a part of fuzzy synthetic decision making.

5. Ranking the strategies (fuzzy number): The result of fuzzy synthetic decision of each alternative is, of course, a fuzzy number. Therefore, it is necessary that the nonfuzzy

0.5 Very low

0 10 15 25 35 40 50 60 65 75 85 90 100 x

low Fair High Very high

1 µ_Ã(x)

ranking method for fuzzy numbers be employed to compare strategies. In other words, the procedure of defuzzification is to locate the Best Nonfuzzy Performance value (BNP).

Methods of such defuzzified fuzzy ranking generally include mean of maximal (MOM), center of area (COA), and α-cut - three methods in all (Zhau & Govind, 1991; Teng &

Tzeng, 1996). To utilize the COA method to find out the BNP is both simple and practical, and there is no need to bring in the preference of any evaluators. For those reasons, the COA method is used in this study.