A Bayesian Reputation System for Virtual Organizations
Abstract: Virtual Organizations (VOs) are an emerging business model in today’s Internet economy. Increased specialization and focusing on an organization’s core competencies requires such novel models to address business opportunities. In a VO, a set of sovereign, geographically dispersed organizations temporarily pool their resources to jointly address a business opportunity. The decision making process determining which potential partners are invited to join the VO is crucial with respect to entire VO’s success. The possibility of a VO partner performing badly during the VO’s operational phase or announcing bankruptcy endangers the investment taken in integrating their processes and infrastructure for the purpose of the VO. A reputation system can provide additional decision support besides the a priori knowledge from quotations and bidding to avoid events such as VO partner replacement by helping to choose reliable partners in the first place. To achieve this, reputation, an objective trust measure, is optimally aggregated from multiple independent trust sources that inherently characterize an organization’s reliability. To allow for the desired predictions of an organization’s future performance, a stochastic modeling approach is chosen. The paper will present a taxonomy of TIs for VO environments, a stochastic model to maintain and aggregate trust sources, so called Trust Indicators, and the inclusion of other subjective measures such as feedback.
Summary (2 min read)
- In today’s business world, commercial relationships become increasingly flexible and are formed on demand whenever a business opportunity emerges.
- Second, the VO manager negotiates with the potential partners, until the required set of VO members is selected, bringing the right set of specialized expertise in the VO.
- In contrast to hard security measures that apply during regular operations between a buyer and supplier such as confidential communication channels or access control measures, reputation based decision support belongs to the class of soft security measures .
- Following up on the above statement about the buyer’s expectation of a supplier, the authors define trust as the subjective probability by which the buyer expects a supplier to behave reliably .
- A feedback mechanism will also be outlined.
2 Trust and Reputation Model
- Many existing trust management approaches, especially in Peer-2-Peer (P2P) environments, root trust solely in feedback given by peers after having conducted a transaction [7, 8].
- Feedback is a highly subjective, relative input that may vary between peers who participated in the same transaction.
- Feedback is still an important input that improves the quality and accuracy of the reputation value aggregating the TI rooted trust values.
2.1 Taxonomy of TIs
- This subsection provides an excerpt of an assembled classification/taxonomy of TIs and their individual modeling.
- Figure 1 shows the top-level TI classes of a taxonomy currently encompassing 146 different TIs.
- TIs modeling the financial trust aspects of an organization, an example is the cash flow quote indicator .
- The distribution assumption, in case of delivery delay that it is exponentially distributed, determines the likelihood distribution P (X|θ) estimating the true parameter θ.
- Each TI incorporates a monotonically increasing weighting function ω(t) > 0 that can implement forgetting of older observations and put emphasis on newer ones.
2.2 Stochastic Model
- The TI’s posterior distributions periodically obtained for each TI after ∆tobs must then be aggregated to a reputation value for this organization.
- The edges denote causal dependencies of the nodes and each node in the BN holds a Conditional Probability Table (CPT) with the probability values for its random variable(s) depending on the parent node’s random variables.
- The bottom layer entails the TI information nodes periodically inputting newly observed data into the BN with the updated TI posterior distribution.
- The same states are held by each middle layer node.
- The information conveyed by an entire probability distribution typically overloads a human being while a computer system further using the reputation distribution for control decisions may benefit from the additional level of detail.
- As mentioned in the beginning, the reputation system relying on observable TIs may still benefit from feedback given by participants of a business transaction who received decision support from the reputation system.
- It has to ensure, e.g. with an established protected session reference that this feedback uniquely relates to the correct transaction and stems from authenticated participants.
- After several experiments with a prototypical implementation of the presented reputation system, that provides reputation values to human requestors, a feedback mechanism accepting feedback in the interval [0, 1] with an associated certainty value of the same interval was contrived, that can be interpreted as a density function as well.
- The feedback hereby applies directly to the overall reputation which more closely aligns to human feedback strategies who typically do not break down feedback to the level of individual TIs but rather follow a gut-feeling approach rating the overall business transaction.
Did you find this useful? Give us your feedback
Related Papers (5)
Frequently Asked Questions (2)
Q1. What have the authors contributed in "A bayesian reputation system for virtual organizations" ?
The paper will present a taxonomy of TIs for VO environments, a stochastic model to maintain and aggregate trust sources, so called Trust Indicators, and the inclusion of other subjective measures such as feedback.
Q2. What have the authors stated for future works in "A bayesian reputation system for virtual organizations" ?
Currently ongoing and future work deals with a set of sound, business domain dependent, configuration sets for bootstrapping a reputation system instance without historic data. The bayesian trust management model itself will be improved by modeling dependencies among TIs themselves on the bottom BN layer. This may potentially lead to cycles in the BN that violates the claim that a directed BN has to be acyclic. Several approaches from BN and graph theory are currently explored, the most promising so far is the junction tree approach, transforming a BN ’ s graph with cycles into a acyclic junction tree [ 16, 17 ].