No abstract

Sequential planning portfolios are very powerful in exploiting the complementary strength of different automated planners. The main challenge of a portfolio planner is to define which base planners to run, to assign the running time for each planner and to decide in what order they should be carried out to optimize a planning metric. Portfolio configurations are usually derived empirically from training benchmarks and remain fixed for an evaluation phase. In this work, we create a per-instance configurable portfolio, which is able to adapt itself to every planning task. The proposed system pre-selects a group of candidate planners using a Pareto-dominance filtering approach and then it decides which planners to include and the time assigned according to predictive models. These models estimate whether a base planner will be able to solve the given problem and, if so, how long it will take. We define different portfolio strategies to combine the knowledge generated by the models. The experimental evaluation shows that the resulting portfolios provide an improvement when compared with non-informed strategies. One of the proposed portfolios was the winner of the Sequential Satisficing Track of the International Planning Competition held in 2014.

Temporal planning is a research discipline that addresses the problem of generating a totally or a partially ordered sequence of actions that transform the environment from some initial state to a desired goal state, while taking into account time constraints and actions' duration. For its ability to describe and address temporal constraints, temporal planning is of critical importance for a wide range of real‐world applications. Predicting the performance of temporal planners can lead to significant improvements in the area, as planners can then be combined in order to boost the performance on a given set of problem instances. This paper investigates the predictability of the state‐of‐the‐art temporal planners by introducing a new set of temporal‐specific features and exploiting them for generating classification and regression empirical performance models (EPMs) of considered planners. EPMs are also tested with regard to their ability to select the most promising planner for efficiently solving a given temporal planning problem. Our extensive empirical analysis indicates that the introduced set of features allows to generate EPMs that can effectively perform algorithm selection, and the use of EPMs is therefore a promising direction for improving the state of the art of temporal planning, hence fostering the use of planning in real‐world applications.

One of the key challenges in automated planning is to define the sources of information that will feed the initial state and goals of each planning task. In many domains, the information comes from company's databases. In other applications, the information is harder to obtain and it is usually partial. In this paper, we will describe an application on travel planning, where the initial state and goals will be obtained by crowdsourcing. Travel planning requires the use of plenty Internet-based resources; some of them are related to human generated opinions on all kinds of matters (e.g. hotels, places to visit, restaurants, ...). We present the OnDroad planner, a system that creates personalized tourist plans using the human generated information gathered from the minube traveling social network. OnDroad proposes an initial tourist guide according to the recommendation of the users profiles and their contacts. In addition, this guide can be continuously updated with newly generated data.

Empirical performance models play an important role in the development of planning portfolios that make a per-domain or per-problem configuration of its search components. Even though such portfolios have shown their power when compared to other systems in current benchmarks, there is no clear evidence that they are capable to differentiate problems (instances) having similar input properties (in terms of objects, goals, etc.) but fairly different runtime for a given planner. In this paper we present a study of empirical performance models that are trained using problems having the same configuration, with the objective of guiding the models to recognize the underlying differences existing among homogeneous problems. In addition we propose a set of new features that boost the prediction capabilities under such scenarios. The results show that the learned models clearly performed over random classifiers, which reinforces the hypothesis that the selection of planners can be done on a per-instance basis when configuring a portfolio.