Centralized Scheduling Solution by FLEXIS. We implemented an agile development process utilizing Scrum methodology. The development was split in two teams, one was implementing the heuristic and the optimization, the other team was implementing the user interfaces and the workflow in the Centralized Planning System. The links and interfaces between the different parts were implemented in a common manner with a floating responsibility between the teams. Additional (weekly) meetings with all the members of the teams and (biweekly) meetings of the team leaders synchronized the activities in the sprints. For the implementation, Java (8 or higher) and Python (3 or higher) was used. These and other technologies that are allowed for the implementation are represented in a flexis internal Tech Radar. All components and code fragments were continuously integrated and continuously deployed to ensure a continuous growth in functionality and continuous testing of the partners in the use cases. Target environment for internal testing was the flexis in house cloud environment. Everything was tested automatically using the unit test framework and Selenium to validate user interfaces.
Centralized Scheduling Solution by FLEXIS. The Centralized Scheduling Solution is implemented as a web-based solution with an internal data model that is populated from a common database via a connector to ensure that all agents are provided with the same data (content and timeliness). The followings figures show the 4+1 diagrams of the Centralized Planning Solution: Figure 23 View – Centralized Scheduling Solution Figure 24 View – Centralized Scheduling Solution – Logic View Figure 25 View – Centralized Scheduling Solution – Development View Figure 26 View – Centralized Scheduling Solution – Process View Figure 27 View – Centralized Scheduling Solution – Use Case View The Centralized Planning Solution is able to schedule different tasks on resources taking into consideration - the process flows to produce a product - the capacities and skills of the resources - the process times of the tasks on the different resources - the shift calendars of the resources - the deadlines and the earliest start date of the orders - the average transportation times of products or components between resources - etc. But the Centralized Planning Solution is not scheduling the transportation tasks in detail, as this can only be done on a very short notice when the locations and the status (free or occupied, loaded or must be loaded, etc.) is known exactly. The Centralized Planning Solution always deals with a production system without disturbances like in an ideal world. When there is a disturbance recognized by the Deviation Agent and the Centralized Planning Solution is forced to make a re-planning, then the Centralized Planning Solution is setting up on the current status of the production which once again can be seen as undisturbed – until the process starts again with a re-planning. The benefit of a Centralized Planning Solution is that the overall production process is optimized, because the Centralized Planning Solution is optimizing the overall make span. The goal of the decentralized planning solutions / agents can be understood as achieving / maintaining the planning result of Centralized Planning Solution as quickly and accurately as possible, even if disruptions occur. The following figure visualizes the data model of the Centralized Planning Solution: Figure 28 Data model of centralized scheduling solution The ins and outs concerning data are (on a high level) visualized in the following figure: Figure 29 Visualization of I/O information The basis of the centralized planning solution is a constructive heu...
