Initial DAI-DSS Prototype – D4.2
Herwig Zeiner, Lucas Paletta, Michael Schneeberger, Gustavo Vieira, Higor Rosse, Rui Fernandes, Sylwia Olbrych, Alexander Nasuta, Magdalena Dienstl, Christian Muck, Rishyank Chevuri
The first part of this documentation deals with the description of the building blocks that form part of the architecture described in deliverable “D4.1 – DAI-DSS Architecture and Initial Documentation and Test Report“. Initially, a quick overview of what the building blocks are and a brief explanation of their general structure are given. We then move on to the integration of the DAI-DSS User Interface (UI) into the system. Taking Industrial Partner CRF's Workload Balance Use Case Scenario as a reference, an overview of the UI in this setting is provided. It describes how the UI offers the possibility of visualising the important components directly linked to decision-making in this example, as well as the added value offered by this building block. The UI is essential for efficient decision-making. It provides the appropriate interface between the decision-maker and the system in order to enable a clear visualisation of the decision-making process. Some images provide an introductory view of the UI in this initial prototype.
Next is the DAI-DSS Orchestrator. The Orchestrator plays a very important and central role in the architecture of this system. Its purpose is to provide coordination in the exchange of information and decisions between the various parts of the system. Like the conductor in an orchestra, the DAI-DSS carries out the decision-making process, and the involved individual services retrieve the relevant data from the knowledge database. This currently takes place in two parallel ways. At a more advanced stage is the Workflow-based Orchestrator. In this orchestrator, a workflow engine offers the possibility of executing workflows, i.e., the execution of tasks that follow a well-defined path leading to the triggering or realisation of an AI service that offers a recommendation for decision-making on issues identified in the business process for which the system is built. For demonstration purposes, a workflow was built in which worker data is retrieved, and the order of a part is requested in the Workload Balance Use Case Scenario. The step-by-step execution of the orchestration is documented in order to elucidate how the orchestration of the building blocks can take place in a real, concrete case from the manufacturing industry in this example.
It presents the DAI-DSS Configurator, a tool designed to enhance decision support systems through efficient configuration and integration frameworks. It explains the configurator's components: the Configuration Framework, which assists in creating decision models and strategies, and the Configuration Integration Framework, which generates system configurations from these models. The report details the prototype's configuration steps, illustrated for clarity. It also allows for microservices and workflow configuration, featuring a user-friendly interface with a wizard for UI components combination.
The DAI-DSS Knowledge Base is highlighted as a central data repository, storing user properties, sensor data, and processed data. It plays a key role in the system's data flow, integrating with the Configurator and using REST API for data retrieval. The report covers the integration of AI services using REST-API endpoints, emphasising ease of access and industry-standard practices like Swagger documentation. It explores various algorithms, including Neural Networks and Decision Trees, and the integration of human factors data for assessing worker resilience. Deployment details of the AI services are discussed, focusing on hosting, management, and future enhancements like data catalogues and decision pattern services for resource allocation and predictive maintenance.
Finally, the report addresses the extension of the DAI-DSS with new services for various industrial use cases. It discusses customising services through conceptual modelling and the advantages of rule-based approaches over machine learning methods. The services aim to optimise resource allocation, improve productivity, and address human factors, thereby enhancing the system's decision-making capabilities in complex industrial scenarios.
Links
Cite as
BibTeX (Download)
@techreport{fairwork_d4.2,
title = {Initial DAI-DSS Prototype - D4.2},
author = {Herwig Zeiner, Lucas Paletta, Michael Schneeberger, Gustavo Vieira, Higor Rosse, Rui Fernandes, Sylwia Olbrych, Alexander Nasuta, Magdalena Dienstl, Christian Muck, Rishyank Chevuri
},
editor = {Gustavo Vieira},
url = {https://fairwork-project.eu/deliverables/D4.2_Initial%20DAI-DSS%20Prototype%20v1.0a-preliminary.pdf},
year = {2023},
date = {2023-12-29},
abstract = {This deliverable, “D4.2 – Initial DAI-DSS Prototype”, is the first reference implementation focusing on the synthesis of fundamental components to establish a shared comprehension of the operational principles behind DAI-DSS. It aims to describe how the different DAI-DSS components fit together, forming a complete and cohesive structure in which the functionalities of the blocks are well-defined, the complementarity of the structure is explored, and the technological component is described in terms of fulfilling its purposes and integrating with the neighbouring structures that are part of the architecture.
The first part of this documentation deals with the description of the building blocks that form part of the architecture described in deliverable “D4.1 – DAI-DSS Architecture and Initial Documentation and Test Report“. Initially, a quick overview of what the building blocks are and a brief explanation of their general structure are given. We then move on to the integration of the DAI-DSS User Interface (UI) into the system. Taking Industrial Partner CRF's Workload Balance Use Case Scenario as a reference, an overview of the UI in this setting is provided. It describes how the UI offers the possibility of visualising the important components directly linked to decision-making in this example, as well as the added value offered by this building block. The UI is essential for efficient decision-making. It provides the appropriate interface between the decision-maker and the system in order to enable a clear visualisation of the decision-making process. Some images provide an introductory view of the UI in this initial prototype.
Next is the DAI-DSS Orchestrator. The Orchestrator plays a very important and central role in the architecture of this system. Its purpose is to provide coordination in the exchange of information and decisions between the various parts of the system. Like the conductor in an orchestra, the DAI-DSS carries out the decision-making process, and the involved individual services retrieve the relevant data from the knowledge database. This currently takes place in two parallel ways. At a more advanced stage is the Workflow-based Orchestrator. In this orchestrator, a workflow engine offers the possibility of executing workflows, i.e., the execution of tasks that follow a well-defined path leading to the triggering or realisation of an AI service that offers a recommendation for decision-making on issues identified in the business process for which the system is built. For demonstration purposes, a workflow was built in which worker data is retrieved, and the order of a part is requested in the Workload Balance Use Case Scenario. The step-by-step execution of the orchestration is documented in order to elucidate how the orchestration of the building blocks can take place in a real, concrete case from the manufacturing industry in this example.
It presents the DAI-DSS Configurator, a tool designed to enhance decision support systems through efficient configuration and integration frameworks. It explains the configurator's components: the Configuration Framework, which assists in creating decision models and strategies, and the Configuration Integration Framework, which generates system configurations from these models. The report details the prototype's configuration steps, illustrated for clarity. It also allows for microservices and workflow configuration, featuring a user-friendly interface with a wizard for UI components combination.
The DAI-DSS Knowledge Base is highlighted as a central data repository, storing user properties, sensor data, and processed data. It plays a key role in the system's data flow, integrating with the Configurator and using REST API for data retrieval. The report covers the integration of AI services using REST-API endpoints, emphasising ease of access and industry-standard practices like Swagger documentation. It explores various algorithms, including Neural Networks and Decision Trees, and the integration of human factors data for assessing worker resilience. Deployment details of the AI services are discussed, focusing on hosting, management, and future enhancements like data catalogues and decision pattern services for resource allocation and predictive maintenance.
Finally, the report addresses the extension of the DAI-DSS with new services for various industrial use cases. It discusses customising services through conceptual modelling and the advantages of rule-based approaches over machine learning methods. The services aim to optimise resource allocation, improve productivity, and address human factors, thereby enhancing the system's decision-making capabilities in complex industrial scenarios.
},
howpublished = {https://fairwork-project.eu/deliverables/D4.2_Initial%20DAI-DSS%20Prototype%20v1.0a-preliminary.pdf},
keywords = {deliverable},
pubstate = {published},
tppubtype = {techreport}
}