Multi-criteria decision making problems exist everywhere. In a complex system, for all involved players, it is inevitable to face task/service selection situations where multiple qualities of tasks/services criteria need to be taken into account. In addition, complex interrelationships between different impact factors and actors need to be understood and traded off. In this paper, we present a tool called I*-Prefer which is represented with annotated NFR/i* framework. This framework uses goal and agent-based preference models to drive these decision making activities. Particularly, we describe the purpose and main features of the tool, and give a brief introduction of the extended i* framework which supports I*-Prefer analysis.
An increasing number of researchers have been working on multi-motive decision models and methods. In order to model and analyze multi-criteria decisions in a systematic way, we have already done some work on how to make decisions based on an annotated NFR/i* framework [1, 2]. In this paper, we present a tool named I*Prefer. Within this tool, we adopt graphical notations of the NFR modeling methods to model the interrelations among different criterion. Then we decide preferences of decision makers by appending numerical annotations to the nodes in the model. Strategy dependency models in i* can be used to represent and evaluate alternative services networking decisions. Algorithms for identifying optimal solutions of the given decision problem are also integrated into the tool.
The structure of this paper is organized as follows: Section 2 presents the original purpose of the tool. Section 3 lists the main features of the tool and its availability and status. Section 4 discusses both the limitations and the future plans. 2
The main purpose of the tool is to model the interrelations among different
criterions. Then we can decide preferences of decision makers by appending
numerical notations to the nodes in the model. I* framework is a widely used strategic
intentional modeling method. Actors’ goals and tasks could be modeled intuitively.
Services selection always involves at least two types of actors, with mutual
dependencies between them. As a result, i* model is a natural fit for modeling
services selection. In order to fulfill the needs to make decision by the preference,
which is the input from domain experts and end users, the tool intends to makes some
extensions to i*: (
Main features and current status of I*-Prefer
I*-Prefer is based on JavaScript, and can be applied in IE5.0 (or higher version). The tool is developed in 2009 which is the version 1.0, the next version is under development.
People can download both its source code and user manual (in Chinese) in the URL[3]: https://sourceforge.net/projects/i-prefer and can also try the tool on line.
The tool can support the methods raised in [2], specifically the main features of the current version are as follows.
(
The tool I*-Prefer supports the quantitative reasoning of actor’s preferences on
soft-goals, and the evaluation of alternative ways to achieve system goals. It can be
used to facilitate automated preference tasks/services selection process and can help
optimize decision making. However, there are also some limitations in the tool: (
In future, we would extend the algorithms to deal with more complex cases, and try to apply the tool into an industrial case. Besides, we would also make the tool more customized and user friendly.