Mixed‐integer linear programming models for batch sterilization of packaged‐foods plants

Batch sterilization with individual retorts is a common mode of operation in many food‐canning plants. Although high‐speed processing with continuous rotary or hydrostatic retort systems is used in very large canning factories, such systems are not economically feasible in the majority of small‐ to medium‐sized canneries. In such canneries, sterilization is carried out in a battery of retorts as a batch process. Although the unloading and reloading operations for each retort are labor intensive, a well‐designed and managed plant can operate with surprising efficiency if it has the optimum number of retorts and scheduling of retort operation. The objective of this research was to present two mathematical models for sterilization scheduling in food‐canning plants. The first model developed is for the case where given amount of different canned food products with specific quality requirements would be sterilized within a minimum plant operation time in an autoclave of given capacity. The second model addresses the problem of maximizing the amount of sterilized products in an autoclave of given capacity for given plant operation time. The developed models were based on mixed‐integer linear programming and incorporated the possibility of simultaneous sterilization. Simultaneous sterilization applies mainly to small canneries with few retorts. In these situations, retorts often operate with only partial loads because of the small lot sizes, and they are severely under‐utilized. In order to demonstrate the feasibility of the mixed‐integer linear programming (MILP) models, several examples involving the sterilization of different products were included in this research. The methodology proposed in this study is of special relevance for small‐ and medium‐sized food‐canning plants that normally work with many different products at the same time.