Simulation of Product-Service-Systems piloting with agent-based models (outlined revision)
Product-service systems (PSS) can improve sustainability in terms of responsibility of providers and intensified customer-provider relationships. Therefore, a PSS approach has been applied on the distribution of solar home systems in rural areas of emerging countries to improve quality and increase confidence in this technology. However, many innovative projects fail at the piloting stage because of insufficient validation before implementation. Validation methods for PSS like an integrated gap model (e.g. identification of gaps between deliverables and customers expectations) or interdisciplinary design reviews show its limits on the interdependences between product and service development. A new approach to test PSS involves agent-based simulation models and is presented within this paper. Among system dynamics and discrete event simulation, an agent-based simulation is able to work with less quantitative data and a more detailed view on individual entities than the alternative approaches. Thus, it is possible to create more realistic simulation models to validate especially resource planning of a PSS. Agent-based simulation is used within the MEVIS (Micro Energy Supply Information System) project to model a network of solar home systems which are controlled by remote monitoring in order to improve service processes, e. g. maintenance and repair operations. The pilot test should be carried out efficiently by the preceding virtual validation in the form of a simulation, which is built in NetLogo, a tool that enables modeling of a multi-agent environment. Within this environment, communicating technical artifacts as well as different groups of actors are mapped as agents. As a result, diagrams of relevant parameters like the overall repair costs or the estimated customer's satisfaction are displayed and the resource planning is visually supported. This helps to push a PSS into a new market, where you have to deal with existing resources that must be used efficiently. The contributions in this paper provide the conception and an exemplary presentation of the simulation run.
Published in: Procedia CIRP, 10.1016/j.procir.2015.02.150, Elsevier