This thesis improves the decision making and management of uncertainty when using Iwao's sequential sampling plan in insect pest management. The objectives of the thesis were addressed in two interrelated parts. First, an approach was developed to select a mean-variance relationship for use in Iwao's sequential sampling plan. Using Monte Carlo simulation, four mean-variance relationships were evaluated on their ability to predict the true variance of the pest population at the decision threshold, a critical component of Iwao's sequential sampling plan. Factors such as the position of the decision threshold along the mean-variance relationship and the number of data points used to estimate the mean and variance played a role in the selection of the relationship. The results of the simulation found that generally that Iwao's mean-variance relationship estimated by , provided the best prediction of the true variance at the decision threshold. Second, uncertainty in the decision threshold was incorporated into Iwao's sequential sampling plan using Monte Carlo simulation. The effect of uncertainty in the decision threshold was to dramatically reduce the accuracy of the sequential sampling plans when compared to sequential sampling plans where the decision threshold was treated as if it was known with certainty. Methods of mitigating the reduced accuracy are discussed. The approaches developed in this thesis provide the pest manager with valuable tools and approaches to improve pest management when using Iwao's sequential sampling plan.
展开▼
