Optimal Wildlife Smuggling Interdiction

Abstract

The illegal trade of wildlife is one of the greatest threats to the survival of many species. Pangolins are claimed to be the most heavily trafficked wild mammals in the world. There is a huge demand for their scales and meat in traditional medicines and food. Wildlife products are trafficked across the world to meet such demands. Smugglers are able to choose the route they take to move these commodities across international borders. Biosecurity resources are placed at ports of entry to intercept smugglers. In this thesis, we describe a model of the relationship between smugglers and biosecurity forces. We begin by assuming that the smuggler picks a path that either minimises travel time, maximises the probability of going undetected, or some combination of these two objectives. Biosecurity agencies allocate resources to ports of entry to maximise their probability of catching the smuggler. However, we assume that the smuggler fully observes any resource changes and thus may alter their path to avoid detection. This results in a dynamic game between the two players. We aim to find the optimal biosecurity resource allocation. However assuming all traffic is sent along a single path from each source is unrealistic. Thus we develop an extension to the model, where the illegal traffic is distributed across the possible paths and only some proportion of the traffic is sent along the optimal path. We describe implementations of two algorithms that find the optimal resource allocation. These solutions lead to insights about the best approaches to interdict illegal trade across ports of entry. We apply our model to a real world network to investigate it in a practical setting. We look at the trafficking of pangolins from Asian and African sources into Vietnam. Resources are allocated at five entry points into Vietnam. We determine the optimal allocations for different smuggler objectives. We consider robust allocations that are good for a majority of the assumed objectives. Spreading resources evenly between three specific ports of entry, out of the five, appears best when we just consider single paths from sources. When we consider multiple paths from the same source, the most robust allocation across smuggler objectives is to spread resources between two specific ports of entry. The ports of entry that are assigned resources in the robust allocations have many more paths into them compared to the other entry points. Thus this robust allocation makes logical sense, and perhaps may apply to situations other than the trafficking of pangolins from Asian and African sources into Vietnam.