Mixed Discrete Continuous Non-Linear Planning through Piecewise Linear Approximation

Denenberg, Elad; Coles, Andrew

doi:10.1609/icaps.v29i1.3469

Cited by 4 publications

(5 citation statements)

References 13 publications

(19 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is in contrast to other related decision making systems that are built on top of additional restrictive assumptions (e.g., the assumption that the state transition function T is piecewise linear (Shin and Davis 2005), polynomial (Cashmore et al 2016) etc.). Our results suggest similar constraint generation approaches can be utilized for effective planning under other planning formalisms that support continuous-time planning Long 2003, 2006;Benton, Coles, and Coles 2012;Scala et al 2016;Micheli and Scala 2019;Denenberg and Coles 2021;Percassi, Scala, and Vallati 2021). Moreover, our theoretical results provide us with some practical insights.…”

Section: Discussion and Related Workmentioning

confidence: 63%

Robust Metric Hybrid Planning in Stochastic Nonlinear Domains Using Mathematical Optimization

Say

2023

ICAPS

View full text Add to dashboard Cite

The deployment of automated planning in safety critical systems has resulted in the need for the development of robust automated planners that can (i) accurately model complex systems under uncertainty, and (ii) provide formal guarantees on the model they act on. In this paper, we introduce a robust automated planner that can represent such stochastic systems with metric specifications and constrained continuous-time nonlinear dynamics over mixed (i.e., real and discrete valued) concurrent action spaces. The planner uses inverse transform sampling to model uncertainty, and has the capability of performing bi-objective optimization to first enforce the constraints of the problem as best as possible, and second optimize the metric of interest. Theoretically, we show that the planner terminates in finite time and provides formal guarantees on its solution. Experimentally, we demonstrate the capability of the planner to robustly control four complex physical systems under uncertainty.

show abstract

Section: Discussion and Related Workmentioning

confidence: 63%

Robust Metric Hybrid Planning in Stochastic Nonlinear Domains Using Mathematical Optimization

Say

2023

ICAPS

View full text Add to dashboard Cite

show abstract

“…DiNo and ENHSP solve all 50 instances relatively quickly and SMTPlan is only able to solve 13 of 50 instances. Figure 4.2 shows Denenberg's [10] results on the linear generator domain with OPTIC ++. When compared to the PDDL 2.1 solver OPTIC++, our planner is not able to solve as many instances as quickly.…”

Section: Discussionmentioning

confidence: 99%

“…ENHSP was run with a discretization of 1. We do not have access to OPTIC ++ and use the results from Denenberg's paper [10]. The goal state in all planning instances is to get the generator to run successfully, i.e.…”

Section: Domain Descriptionmentioning

confidence: 99%

Heuristic Planning for Continuous Systems in Hybrid Temporal Situation Calculus

Mathew

2024

Preprint

View full text Add to dashboard Cite

<p>Given a description of domain and its dynamics, temporal numeric planning attempts to find a sequence of actions that satisfies a given set of constraints for a dynamical system. Current planners operate on grounded transition systems and discretized representations of the domain which lead to poor scalability. Furthermore, given the problem’s difficulty, most modern planners restrict their capabilities to a subset of hybrid domains, e.g. support for only polynomial evolution of numeric state variables and linear action conditions. To address these concerns, we present a lifted planner, NEAT (Non-linEAr Temporal) Planner, that utilizes a logical description of the domain described in Hybrid Temporal Situation Calculus. Furthermore, we develop AMPLEX, an interface to AMPL and several non-linear programming solvers, which allows us handle several non-linear functions. We also present a novel non-linear programming based heuristic to improve scalability. Lastly, we perform a detailed comparison between current state-of-the-art solvers and our planner.</p>

show abstract

“…For the updated problem, the bounds are sorted, and processes generated for each adjacent pair of bounds, each with its own r ub,lb value. This collection of processes uses a "stacked" model to generate a combined linear effect (Denenberg and Coles 2018;2019).…”

Section: Refining the Linearizationmentioning

confidence: 99%

Personalized Medication and Activity Planning in PDDL+

Alaboud

Coles

2019

ICAPS

View full text Add to dashboard Cite

The emergence of planners capable of reasoning with continuous dynamics, as expressed in PDDL+, has increased the range of problems that fall within the capabilities of PDDL planners. One such problem is planning patients’ activities and medication regimes, considering non-linear medication pharmacokinetics. In this paper we explore the application of contemporary PDDL+ planners to this problem. To address their performance limitations, we present a linearize–validate cycle; tasks are solved by iterative refinement of a linear approximation of the domain, solved by a linear planner, then validated at each stage against the full non-linear semantics. In doing this we allow this domain to fall within the capabilities of current planners; and in our evaluation we use OPTIC to demonstrate this.

show abstract

Mixed Discrete Continuous Non-Linear Planning through Piecewise Linear Approximation

Cited by 4 publications

References 13 publications

Robust Metric Hybrid Planning in Stochastic Nonlinear Domains Using Mathematical Optimization

Robust Metric Hybrid Planning in Stochastic Nonlinear Domains Using Mathematical Optimization

Heuristic Planning for Continuous Systems in Hybrid Temporal Situation Calculus

Personalized Medication and Activity Planning in PDDL+

Contact Info

Product

Resources

About