A Markov Decision Process Model for Strategic Decision Making in Sailboat Racing

Ferguson, Daniel; Elinas, Pantelis

doi:10.1007/978-3-642-21043-3_14

Cited by 9 publications

(5 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They are often used in regattas to help sailors to take the best actions given certain weather conditions. Such approaches use Markov decision processes 31,32 and Markov chains 9 for estimating the trajectory that is optimum, helping human sailors to take their decisions.…”

Section: Work Dealing With the Upwind Situationmentioning

confidence: 99%

A gain-scheduling control strategy and short-term path optimization with genetic algorithm for autonomous navigation of a sailboat robot

Santos

Gonçalves

2019

International Journal of Advanced Robotic Systems

View full text Add to dashboard Cite

The development of a navigation system for autonomous robotic sailing is a particularly challenging task since the sailboat robot uses unpredictable wind forces for its propulsion besides working in a highly nonlinear and harsh environment, the water. Toward solving the problems that appear in this kind of environment, we propose a navigation system which allows the sailboat to reach any desired target points in its working environment. This navigation system consists of a low-level heading controller and a short-term path planner for situations against the wind. For the low-level heading controller, a gain-scheduling proportional-integral (GS-PI) controller is shown to better describe the nonlinearities inherent to the sailboat movement. The gain-scheduling-PI consists of a table that contains the best control parameters that are learned/ defined for a particular maneuver and perform the scheduling according to each situation. The idea is to design specialized controllers which meet the specific control objectives of each application. For achieving short-term path-planned targets, a new approach for optimization of the tacking maneuvering to reach targets against the wind is also proposed. This method takes into account two tacking parameters: the side distance available for the maneuvering and the desired sailboat heading when tacking. An optimization method based on genetic algorithm is used in order to find satisfactory upwind paths. Results of various experiments verify the validity and robustness of the developed methods and navigation system.

show abstract

Section: Work Dealing With the Upwind Situationmentioning

confidence: 99%

A gain-scheduling control strategy and short-term path optimization with genetic algorithm for autonomous navigation of a sailboat robot

Santos

Gonçalves

2019

International Journal of Advanced Robotic Systems

View full text Add to dashboard Cite

show abstract

“…DP has also been extensively investigated for decision-making in different applications, including energy (Clement-Nyns et al, 2010), logistics (Hall and Potts, 2003), medicine (Sahinidis, 2004). Markov Chains have been used also by Ferguson and Elinas (2011), in a study focussed on finding optimal routes for inshore racing, in presence of landmasses influencing the wind. In this case, the environment is considered fully observable (i.e.…”

Section: State Of the Art On Yacht Racing Strategy And Contribution Omentioning

confidence: 99%

A real-time strategy-decision program for sailing yacht races

Tagliaferri

Viola

2017

Ocean Engineering

View full text Add to dashboard Cite

A B S T R A C TOptimal decisions for a skipper competing in a match race depend on a number of factors, including wind speed and direction variations, behaviour of the opponent, sea state, currents, racing rules. Expert sailors are able to combine observations on these various factors and process them to take optimal decisions. This study presents an attempt at emulating this decision process through a computer code that can be used in real time to advise on race strategy. The novelty of the proposed method consists in combining various approaches for the multiple factors affecting the decision process.The wind variability is modelled with the use of neural networks, to produce a short-term wind forecast. The willingness of the sailor to risk is modelled using coherent-risk measures. Experimental results are used to quantify the loss of speed due to the presence of a nearby opponent. Finally, all these factors are combined through dynamic programming to compute an optimal course, based also on information on the current and yachts performance. The program is tested modelling the last 13 races of the 34th America's Cup, and results show that the route computed is close to the shortest possible route computed assuming perfect knowledge of sea conditions.

show abstract

“…Other studies not directly related to the America's Cup have tackled the problem of decisionmaking for sailors. Ferguson and Elinas (2011) propose a simple Markov decision model, where at all times the sailor has only two options, "do nothing" or "tack". The work developed in this study, focussed on inshore racing, includes a VPP and a model for wind flow around landmasses.…”

Section: Yacht Racing Strategymentioning

confidence: 99%

Development of a Routing Software for Inshore Match Races

Tagliaferri

Viola

2016

Day 2 Sat, March 19, 2016

View full text Add to dashboard Cite

Yacht races are won by good sailors racing fast boats. A good skipper takes decisions at key moments of the race based on the anticipated wind behavior and on his position on the racing area and with respect to the competitors. His aim is generally to complete the race before all his opponents, or, when this is not possible, to perform better than some of them. In the past two decades some methods have been proposed to compute optimal strategies for a yacht race. Those strategies are aimed at minimizing the expected time needed to complete the race and are based on the assumption that the faster a yacht, the higher the number of races that it will win (and opponents that it will defeat). In a match race, however, only two yachts are competing. A skipper’s aim is therefore to complete the race before his opponent rather than completing the race in the shortest possible time. This means that being on average faster may not necessarily mean winning the majority of races. This papers present the development of software to compute a sailing strategy for a match race that can defeat an opponent who is following a fixed strategy that minimizes the expected time of completion of the race. The proposed method includes two novel aspects in the strategy computation: A short-term wind forecast, based on an Artificial Neural Network (ANN) model, is performed in real time during the race using the wind measurements collected on board. Depending on the relative position with respect to the opponent, decisions with different levels of risk aversion are computed. The risk attitude is modeled using Coherent Risk Measures. The software is tested in a number of simulated races. The results confirm that maximizing the probability of winning a match race does not necessarily correspond to minimizing the expected time needed to complete the race.

show abstract

A Markov Decision Process Model for Strategic Decision Making in Sailboat Racing

Cited by 9 publications

References 9 publications

A gain-scheduling control strategy and short-term path optimization with genetic algorithm for autonomous navigation of a sailboat robot

A gain-scheduling control strategy and short-term path optimization with genetic algorithm for autonomous navigation of a sailboat robot

A real-time strategy-decision program for sailing yacht races

Development of a Routing Software for Inshore Match Races

Contact Info

Product

Resources

About