Several challenges have been identified in the dairy bulk powder commodity sector due to the nature of this particular manufacturing environment. Internal and external factors compromise the efficiency of the production process and must be mutually addressed. The impact of seasonality and the frequency of deliveries performed by milk suppliers are evident in dairy manufacturing since they restrict the raw material availability and cause bottlenecks at the initial manufacturing stages due to the variability of the raw material supplied. A short raw material lifespan restricts the time for intermediate storage. In discrete manufacturing processes, for example, parts can be stored for extended periods while in dairy manufacturing, the raw material has a limited time to be processed due to a high level of perishability. Therefore, a fast rate in processing Work-in-Process (WIP) is required owing to the limited storage capacity. Constant interruptions caused by Cleaning in-Place (CIP) cycles are critical since they compromise the system processing capacity and also must comply with regulatory authorities such as Food and Drug Administration (FDA) and Good Manufacturing Practices (GMP) to reduce contamination risks. Therefore, operations must be adapted to increase the response during such unforeseen events. The complexity is potentially increased when available resources are poorly managed due to the lack of visibility, and inefficient support decision tools intensify this scenario as observed by the increase of WIP levels at distinct manufacturing stages without adding value to the final product processing. This research conducts an investigation incorporating the internal and external factors in the sector investigated on a framework designed to simulate this environment entitled ePMPS (e-Pilot Manufacturing Plant Simulation). The innovation of the architecture proposed is the ability to simulate several scenarios through Software-as-a-Service (SaaS) where inputs are received in a structured file and the output is presented on a multi-device dashboard. The dairy manufacturing environment is generalised into a three-stage production flow where the main Production Control Strategies (PCSs) traditionally implemented in discrete manufacturing are investigated and replicated in this environment. The main pull-type systems were mathematically modelled in a Markov Decision Process (MDP) considering the impacts of CIP cycle times and supply variability. In addition, the requirements for implementing production policies such as CONstant Work-in-Process (CONWIP), Kanban Control Strategy (KCS), and Hybrid CONWIP/KCS are explored and compared to two distinct seasons. By examining the trade-off between conflicting objectives: maximising Service Level Agreement (SLA) and minimising WIP levels, it is possible to observe that this sector requires specific WIP levels at each manufacturing stage. The results presented by KCS demonstrated an efficient strategy in absorbing the variability caused by supply deliveries during peak season since a systematic volume authorisation controls the excess of material within the system. During off-seasons or specific periods of the year under extra production capacity, an optimised production plan according to the volume supplied is more appropriate. Distinct strategies are required for distinct seasons and the results demonstrated have provided a fundamental basis to support decision-makers in addressing the challenges faced by this sector.