Tesis doctoral de Carlos David Mérida Campos
Multi-party negotiation for complex goods or tasks are often driven by a request for proposal (rfp) protocol. Rfp based systems are those in which an entity submits an invitation (a request for proposal) for providers of a product or service to bid on the right to supply that product or service. In the real world, such requests are often addressed by groups or consortia of agents (coalitions) combining their skills in order to fulfill a need effectively. Requests can also often repeat or change characteristics over time forcing agents to adapt their bids and consortia — leading to competition between dynamically changing groups within the population. This thesis sets out to model such environments as coalitional multi-agent systems and in particular, to understand rfp events as dynamic processes. the model studied considers heterogeneous populations of self interested agents that iteratively reconfigure proposed solutions to fit dynamic demand. Social networks between agents are also considered to capture the possible sight and/or computational limitations of agents in large scale systems. In particular, the objective of the work is to explore the dynamics of a coalitional framework of a rfp negotiation protocol and thereby answer the following research questions: how can agents exploit incomplete data over time?, How can agents reuse coalitional structures in large-scale coalition formation systems? And how can agents reuse social structures in large-scale coalition formation systems?. To address these questions, the rfp negotiation model has been studied from different perspectives, including scarcity of information, heterogeneity of capabilities or requirements, different types of social awareness, the effect of social topologies and simultaneity of requests. the model demonstrates interesting stability and performance characteristics indicating that negotiation of complex goods in dynamic environments can be effectively managed using the rfp protocol, and also that dynamic coalition formation can be directly supported by social structures to be successfully adapted to contextual changes.
Datos académicos de la tesis doctoral «Dynamic coalition formation in iterative request for proposal environments«
- Título de la tesis: Dynamic coalition formation in iterative request for proposal environments
- Autor: Carlos David Mérida Campos
- Universidad: Politécnica de catalunya
- Fecha de lectura de la tesis: 12/06/2009
Dirección y tribunal
- Director de la tesis
- Steven Nicolas Willmott
- Tribunal
- Presidente del tribunal: boi Faltings
- onn Shehory (vocal)
- dirk sascha Ossowski (vocal)
- mike Luck (vocal)