Abstract: This paper was motivated by the discussion held at the 2nd YMCDA meeting on the theme “What is a constructive approach in MCDA?”. On the one hand, it intends to organise some thoughts on the differences between normative, descriptive, prescriptive and constructive approaches, as well as their relation (or lack of relation) with MCDA methods. On the other hand, it intends to discuss many of the questions that occur to MCDA researchers (not necessarily the younger ones!) about such issues.

The purpose of this work is to improve understanding of existing and proposed decision systems, ideally to improve the design of future systems. A "decision system " is defined as a collection of information-processing components-- often involving humans and automation (e.g., computers)-- that interact towards a common set of objectives. Since a key issue in the design of decision systems is the division of work between humans and machines (a task known as "function allocation"), this report is primarily intended to help designers incorporate automation more appropriately within these systems. This report does not provide a design methodology, but introduces a way to qualitatively analyze potential designs early in the system design process. A novel analytical framework is presented, based on the concept of "semi-Structured " decision processes. It is believed that many decisions involve both well-defined "Structured " parts (e.g., formal procedures, traditional algorithms) and ill-defined "Unstructured " parts (e.g., intuition, judgement, neural networks) that interact in a known manner. While Structured processes are often desired because they fully prescribe how a future decision (during "operation") will be made, they are limited by what is explicitly understood prior to operation. A system designer who incorporates Unstructured processes into

A necessary precursor to strategic choice is the selection of a strategic choice evaluation method — metachoice. Unfortunately, managers are unlikely to find much guidance on how to conduct metachoice analysis in the strategy literature. In light of this problem, a template is presented based on two questions that must be addressed by those engaged in strategic analysis generally and metachoice specifically: (1) what are the organization’s strategic goals? (2) how willing are they to monetize their predictions and valuations? Four resulting choice methods are presented and discussed: Discounted Cash Flow Analysis (including variants based on real options), Profitability Analysis,

The Allais Paradox seems as fresh today as it was when first posed some 40 years ago. It has been discussed in the intervening decades and led to many suggestions for modifying, extending or abandoning standard subjective expected utility theory. It has also led to much discussion of the distinction between normative and descriptive models and their rle in prescriptive decision support. It seems appropriate in a conference entitled Decision Making: towards the 21st Century to reflect on these developments, particularly in relation to recent extensions to utility theory and other mathematical models of preferences, and consider, in particular, their possible implication for prescriptive decision analysis. Keywords: decision analysis; descriptive and normative models; modelling of preference; prescriptive analyses; subjective expected utility model. 1.Background The International Colloquium on Risk held in Paris in 1952 was a significant event in the history of decision theory. Many o...

40 41 42 43 44 45 46 47 48 49 50

40 41 42 43 44 45 46 47 48 49

Abstract: This paper addresses the role of Personal Digital Assistants (PDAs) in electronic meetings. Several real-world scenarios of PDA usage in meetings are defined using a pattern language. Anchored on these scenarios, we propose an upper-layer meeting middleware which addresses three major goals: defining a common architecture and set of components for meeting systems; standardizing the meeting memory and process data structures commonly managed by electronic meetings; and supporting XML-based interoperability between these components. The patterns, architecture and components were validated through their adoption in three applications, developed by different teams and covering quite different domains. The applications, encompassing several meeting patterns and adopting multifaceted combinations of the upper-layer components, demonstrate the high level of interoperability supported by the proposed upper-layer middleware.

40 41 42 43 44 45 46 47 48 49

Any statistical analysis or decision analysis contains numerical inputs of which we are unsure. Some of our uncertainty arises from physical randomness which we can model in various ways, ideally through probability. Some relates to judgmental estimates of qualities about which we may be unsure in many different respects. There are other uncertainties involved, however: some relates to ambiguity and imprecision of meaning; some relate to lack of clarity in the objectives which the analysis seeks to meet; some relate to the numerical accuracy of calculations. Suppose, for instance, that we need to consider the probability of a conservative government being returned at the next election. The probability, a subjective probability, itself is a number that the statistician or decision maker has to judge. In some ways it may be related to 'hard' data, but there will always be an element of judgement in setting its value. But one might also ask: what is meant by a 'conservative' government; ...

Abstract not found