An Orchestrated Multi-issue Negotiation with Normal Distributions for Service Selection

In QoS-aware Service-Based Applications (SBAs) composed of dierent and autonomous services, QoS preferences are expressed by consumers as end-to-end quality requirements on the whole application. Software agents negotiation can be adopted to select the component services for which an agreement on the provided QoS values is reached with the SBA consumer. In the present work, we adopt a hybrid iterative negotiation mechanism when end-to-end requirements cannot be mapped to local requirements for each service, so consumers have to negotiate in a coordinated way with the providers of dievent services. We show that in composition of services negotiation is inherently multi-issue also when a single issue is considered for the entire application. The use of normal distributions used in the adopted negotiation mechanism allows to model single-issue and multi-issue negotiation within the same negotiation framework in terms of adopted concession strategy and utility.