An integrated simulation and AHP approach to vegetable grafting operation design

•A simulation model is proposed for vegetable grafting operation design.•Three automation levels (manual, semi-, fully-automated) are considered for grafting operation.•AHP is employed to address multi-criteria decisions in grafting operation design.•We allow propagators to select an appropriate grafting operation design based on the conditions.

For vegetable seedling propagators, integrating grafting technology into their propagation operations is critical to keeping or expanding market share in the near future. In this paper, an integrated discrete event simulation and Analytic Hierarchy Process (AHP) approach is proposed to help vegetable seedling propagators design grafting operation. The proposed approach consists of four steps: (1) defining performance criteria and factors (i.e. system alternative parameters and noise factors), (2) identifying significant factors via Design of Experiment (DOE), (3) evaluating system alternatives, and (4) AHP. For steps 2 and 3, a generic propagation simulator is developed with the focus on specific grafting operations (e.g. creating rootstocks, scions and grafted seedlings) while considering biological factors (e.g. seed disease and disease infection). Both classic and fuzzy AHP methods are adopted for addressing multiple criteria (e.g. variable cost, grafting throughput time, total capital expenses, resource utilization and percentage of order fulfilled in time with acceptable quality) of decisions. To address the imprecise ranking led by utilizing sample means of alternative performance data in pairwise comparison, a Best Alternative Search (BAS) procedure is proposed for AHP by considering bounds of confidence intervals in ranking alternatives. In the experiments, six system alternatives involving three automation levels (e.g. manual, semi- and fully-automated grafting) and eight scenarios are applied to a large-scale seedling propagator located in North America. Results demonstrate that (1) classic AHP produces the similar trend as fuzzy AHP, (2) the proposed BAS procedure can ensure the ranking accuracy of AHP, and (3) the proposed approach can be successfully used by vegetable seedling propagators to support the design of a grafting operation.