![]() One of the operations – PYTHON – implements cluster-level interoperability with Python (installed on the same computer as InterSystems IRIS). Operations and processes included on isc.py.test.Production production (see Figure 2 above) are all based on ML Toolkit extensions for InterSystems IRIS, in particular, on community Python Gateway component. Their descriptions can be found in Table 1 above, we will just provide examples of each variable materialized in InterSystems IRIS:įigure 3: InterSystems IRIS visual process editor with steps in a process workflow InterSystems IRIS provides the following state variables: production ( more ), operation ( more ), process ( more ), and step ( more ). 2.3 State VariablesĮach of the major components of the prototype that we explore, InterSystems IRIS and NetLogo, has its own state variables. 2.2 EntitiesĮntities operating at cluster and factory scales, coincide with those scales: factory cluster – a set of factories connected via their inputs/outputs, and factory – a factory servicing its flow of incoming customer orders and providing its output as input to the other factories.įactory operation scale is shared by two entities: agents – robots, production machinery, warehouses (loading docks and storages) and the global entity – a set of all the global variables used in NetLogo models implementing both reality-mappable and abstract attributes. Spatial coordinates (plus “world wrapping” options) allow representing virtually any factory floor layout. Spatial dimension is captured using NetLogo spatial (2D) coordinates in combination with agents implemented as “patches”. ![]() Both allow pacing execution speed to fit virtually any use case: from reproducing actual time scale to going at accelerated pace (e.g., for simulation as part of a forecasting mechanism, etc.). Temporal dimension is captured through both hang and wait commands respectively in InterSystems IRIS and NetLogo. We distinguish three scales: factory cluster – where interconnected factories are regarded as a single unit, factory – where a factory is considered a single unit, and factory operation – where factory parts are modeled individually. Table 1: Entities, state variables, and scales 2.1 Scales The three NetLogo interconnected robotic factory models participating in the prototype can also run outside of the data platform environment of InterSystems IRIS – thus preserving prototype’s agent-based simulation functionality but losing in-platform orchestration controls as well as cluster-level descriptive and predictive analytics. In this text, we focus on the effects from orchestrating agent-based factory cluster simulation via an all-embracing universal data platform. Again, for research on the effects from various in-platform implementations of factory cluster simulation, we would refer reader to appropriate studies. The advantages of in-platform implementation of factory cluster simulations have been materialized more recently due to the evolution of computers and software making feasible parallel computations, near-real-time integration exchanges and seamless use of a full spectrum of modeling toolsets (e.g., Ng et al., 2011 ). In this text, we focus on functional benefits from doing factory cluster simulation using agent-based approach. For research on the effects from various configurations of factory clusters, we would refer the reader to those studies. The high potential that inter-factory linkages simulation has in integrated production network setting, was established in numerous academic and applied studies (e.g., Ferdows and Carabetta, 2006 ) since long ago. For the underlying prototype, NetLogo suite was used to do factory agent-based simulation (re-using “Robotic Factory” model ) while InterSystems IRIS data platform was used for NetLogo orchestration and factory/cluster end-to-end simulation. In this paper we prototype and explore how multiple agent-based models of robotic factories connected to other robotic factories (represented by their respective models) can be orchestrated using an all-purpose data platform – thereby simulating descriptive and predictive properties of a group of factories (a factory cluster). * In-Platform Agent-Based Simulation of a Connected Factory ClusterĪuthor: Sergey Lukyanchikov, InterSystemsĪ live demo recording can be found here: 1.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |