Goal recognition design

Abstract

We propose a new problem we refer to as goal recognition design (grd), in which we take a domain theory and a set of goals and ask the following questions: to what extent do the actions performed by an agent within the model reveal its objective, and what is the best way to modify a model so that any agent acting in the model reveals its objective as early as possible. Our contribution is the introduction of a new measure we call worst case distinctiveness (wed) with which we assess a grd model. The wed represents the maximal length of a prefix of an optimal path an agent may take within a system before it becomes clear at which goal it is aiming. To model and solve the grd problem we choose to use the models and tools from the closely related field of automated planning. We present two methods for calculating the wed of a grd model, one of which is based on a novel compilation to a classical planning problem. We then propose a way to reduce the wed of a model by limiting the set of available actions an agent can perform and provide a method for calculating the optimal set of actions to be removed from the model. Our empirical evaluation shows the proposed solution to be effective in computing and minimizing wed.