Task: Design Trust Measurement Infrastructure

Design the elements of the system that allow the derivation of experiences.

Purpose

Provide the infrastructure required to derive trust values.

Relationships

Parent Practices	Trust-based Interaction Design
Roles	Primary Performer: Architect	Additional Performers: Developer
Inputs	Mandatory: [Multi-Agent System Architecture] [Requirements Model]	Optional: [Agent Architecture] [System Environment Description] Trust Model
Outputs	[Agent Architecture] [Multi-Agent System Architecture]

Main Description

In order to measure the trust value, contracts have to be established and the results of interactions have to be compared to these contracts. For this purpose, the desired and the actual result of interactions need to be determined and transformed into an experience. Depending on the interaction, the contract can either be stated explicitly or even negotiated in a separate interaction or be defined implicitly.

In any case, experiences have to be stored in the system and made available to the trust model. The trust model will have to be able to access the experiences and create a trust value from them. The value will then have to be available to the agent or a reputation manager.

Existing patterns for this infrastructure should be employed whenever possible.

Steps

Specify the contents of contracts

This step has to be performed for each interaction with different contracts. The designer has to make sure that all relevant aspects of the interaction are captured in the contract. Depending on the actual interaction, this includes prices, delivery deadlines, maximum response times, etc.

Define the data stored in experiences

Experiences can store pre-processed data, such as deviations from the contract, or raw data, e.g., the contract and the actual result of the experience. Depending on the trust model used, only relevant information must be stored. However, discarding information can also mean that the system is not flexible w.r.t. later changes in the trust model. The designer has to make a conscious and well documented decision which data is kept and which pre-processing steps have to take place.

Design interfaces to the trust model and the agents

In order to make the data available to the trust model and to the agents, appropriate interfaces have to be defined. If a reputation system is used, make sure that it can also access the trust values or it automatically receives new trust values (pull or push).

More Information

Supporting Materials	The Trust Lifecycle
Whitepapers	Anders et al. - Patterns to Measure and Utilize Trust in Multi-Agent Systems

This material is made available under the Creative Commons—Attribution-ShareAlike License v3.0.

Parts of this material is based on the Eclipse Process Framework Practice Library, distributed under the Eclipse Public License V1.0.