PA—Precision Agriculture: Optimisation of Lily Microbulb Multiplication Operations

A discrete simulation model was developed to simulate daily processes and output for lily microbulb multiplication. The model employed lognormal distribution to describe the probabilistic behaviour of process factors: entry vessel quantity, single-vessel processing time and multiplication rate. In order to determine the optimum operation conditions for the multiplication process, the discrete model was used to generate predicted values for a three-variable Box–Behnken design of response surface analysis. Second-order polynomial equations were then regressed to predict the finished vessel quantity and total processing time, as affected by the process factors. Predicted values from the second-order polynomial equations and model simulations were also compared with the experimental data. Canonical analysis revealed that the response surface of daily finished vessel quantity and total processing time were both saddle surfaces that contained neither global maximum nor global minimum. A constraint optimization of process variables could be identified from contour plots of the response surfaces, which suggested more productive process in limited daily processing time.