Soo-Hwa is a Lab Researcher working on the Brokered Additive Manufacturing (BAM) project specifically investigating “machines with memory” using agent based modelling. She is currently finishing her PhD at the University of Bath. Her project, “Meshless Additive Manufacturing”, is to create mechanically tunable triply-periodic minimal surface (TPMS) lattices using implicit-based modelling.
The “machines with memory” project focuses on implementing adaptive memory algorithms within the machine agents to investigate the effects on job selection. The BAM models mimic realistic scenarios where each job has different characteristics, such as print time, filament type, etc., as well as the dynamics of different job demand profiles over time. Additionally, machine agents have realistic operation time penalties, such as a day-night scheme, changing filament, printing time, post processing time, etc. By optimally selecting jobs based on their characteristics and adapting to the dynamics this can potentially increase the overall productivity and reduce bandwidth.
The investigated approach starts with the machine agents sampling available jobs to survey the job characteristics. Using this information, the agents then can adaptively select for a subset of the sampled jobs, i.e. a batch, to complete. The batch has its own selection process in which it selects jobs based on minimising time penalties incurred by the machine.