Game Theoretic Fish Patrol Schedule Model

Coast Guard boat

Game Theoretic Fish Patrol Schedule Model

Motivation

Marine fisheries are acknowledged to be some of the most important food resources for countries around the world. However, the issue of fishery sustainability has now become a key concern around the world. As reported by World Wild Fund for Nature (WWF), cod are currently at risk from overfishing in the UK, Canada and most other Atlantic countries. Global cod catch has suffered a 70% drop over the last 30 years, and if this trend continues, the world’s cod stocks will disappear in 15 years.

Illegal, unreported, and unregulated (IUU) fishing is one of the major threats to the sustainability of ocean fish resources. As estimated by National Oceanic and Atmospheric Administration (NOAA), IUU fishing produces between 11 and 26 million tons of seafood annually, representing as much as 40 percent of the total catch in some fisheries. The driver behind IUU fishing is high economic profit and low chance of seizure.

It is impossible to maintain a 24/7 presence to prevent IUU fishing everywhere due to the limited asset patrolling resources. Hence the allocation of the patrolling resources becomes a key challenge for security agencies like USCG. Research within this project aims to address the problem of deriving accurate patrol schedules for the US Coast Guard. To achieve this, our aim is to develop fast and effective techniques to solve security games defined over continuous spaces, where different types of defenders and attackers might co-exist and potentially coordinate their behavior.

Game Theoretic Fish Patrol Schedule Model Overview

The Game Theoretic Fish Patrol Schedule Model casts the interaction between the USCG (defender) and the illegal fishing boats, i.e., lanchas (adversaries), as a repeated Stackelberg game. Real-world data provided by the USCG is used to estimate the parameters of the game, including the behavorial model of the Lancha adversaries. The Stackelberg game can be solved using the MIDAS algorithm which computes the defender’s strategy, i.e., a randomized patrolling strategy. The defender strategy is then used to generate daily patrol schedules for the USCG assets. Executing these patrol schedules produces more Lancha data which allows the entire process to be repeated.

  • Aerial asset

    Airplane

  • Will Haskell

    Will Haskell next to airplane

  • Fishing vessel

    Fishing vessel

  • Coast Guard short range boat

    Coast Guard short range boat

  • Coast Guard patrol boat

    Coast Guard patrol boat

Disclaimer

The views expressed in this article are those of the author and DO NOT necessarily reflect the views of the USCG, DHS, or the U.S. Federal Government or any Authorized Representative thereof. The U.S. Government does not approve/ endorse the contents of this article, therefore the article shall not be used for advertising or endorsement purposes. The U.S. Government assumes NO liability for its contents or use thereof. Moreover, the information contained within this article is NOT OFFICIAL U.S. Government policy and cannot be construed as official in any way. Also, the reference herein to any specific commercial entity, products, process, or service by trade name, trademark, manufacturer, or otherwise, does NOT necessarily constitute or imply its approval, endorsement, or recommendation by the U.S. Government.

Coast Guard officer in a fishing boat

Milind Tambe

Matthew Brown

Debarun Kar

Yundi Qian

William Haskell

Albert Xin Jiang

 

International Joint Conference on Artificial Intelligence (IJCAI), 2015

Fei Fang, Peter Stone, Milind Tambe
When Security Games Go Green: Designing Defender Strategies to Prevent Poaching and Illegal Fishing
  • Coast Guard Atlantic Area
  • Coast Guard Pacific Area
  • Coast Guard Headquarters
  • Coast Guard District One
  • Coast Guard District Eight
  • Coast Guard District Eleven
  • Coast Guard Research and Development Center
  • DHS S&T, Office of University Programs