Repo: HERE
This manufacturing facility creates three types of products (P1, P2, P3), each requiring at least one or more component (C1, C2, C3) for its creation. Each product has a specific workstation associated with it, which is solely responsible for that item. Before assembly, every workstation must wait for the required components, using a buffer zone with a maximum capacity of two for each assembly piece. Two inspectors check the components before reaching the buffer and place them in the specified workstation if the buffer zone is not full. Inspector 1 chooses the buffer zone with the least number of components waiting in the queue. If there's a tie, Workstation 1, 2, then 3 is favored. A simulation is a practical way to determine the ideal workflow, given the complexity of the system states. It initially appears likely that Workstation 1 will have a higher production rate than Workstation 2 or 3, as it only requires one component, without considering each Workstation’s processing time.
The goal of this simulation is to identify the best behavior for Inspector 1 and to determine the best priority of each Workstation for distributing Component 1 to maximize production. To identify the ideal workflow, the criteria for Inspector 1 to distribute Component 1 must be changed, along with the priority of each Workstation when a tie occurs.
Products:
Three potential products (P1, P2, P3) are created at the facility.**
Components:
Each product requires different components (C1, C2, C3) for production. The supply of components is unlimited and will not slow down the production process.
Inspectors:
Two inspectors are responsible for inspecting components. Inspector 1 is in charge of C1, while Inspector 2 handles C2 and C3. The inspectors place the components into the buffer zones adjacent to the workstations.
Workstations:
Each product (P1, P2, P3) has an associated workstation (W1, W2, W3) where it is assembled.**
Buffer Zone:
The component queue for the workstations. Each workstation can hold a maximum of two components each, and five buffer zones exist for each product (P1, P2, P3).**
The rate of production is limited by the inspectors, workstations, and buffer zones. The time each inspector takes to inspect each component is critical. Although the workstations are important, their production rate is fixed and depends on the supply of components. If components are readily available in the buffer, the products will be produced at the rate defined for each workstation. The supply of components is unlimited, hence not considered an essential feature or bottleneck for the system.
Two key assumptions are made in the simulation. The first assumption is regarding the behavior of Inspector 2. As Inspector 2 handles two components, it's assumed that the selection between Component 2 or 3 is either random or based on a defined interval. Secondly, if an Inspector is blocked from placing components as the buffer zones are full, it's assumed that Inspector 2 will move onto another needed component or hold onto the component without starting a new inspection if all buffer zones are filled.
MATLAB was chosen as the scripting language to develop the simulation model. MATLAB has a discrete-event simulation framework (Simulink) that can handle the translation of the model into a simulation. It is a common programming language used in machine learning, simulation, and designs that require multiple iterations. The distributions for the inspection times, component selection, and workstation processing times are randomly sampled to create a distinct simulation of random events. Scripts are used to process these events and output statistical information about the efficiency and quality of the choices. Each event is processed as they occur, with multiple checks for buffer limitations and workstation production. Inspector idle times for when buffer queues are full are also considered. The results from the simulation are validated to ensure they are acceptable.
…return to projects.ibrahimsaid.ca or ibrahimsaid.ca.