Planning as Goal Recognition: Deriving Heuristics from Intention Models -- Extended Version

Giacomo Rosa; Jean Honorio; Nir Lipovetzky; Sebastian Sardina

cs.AIMar 16, 2026

Planning as Goal Recognition: Deriving Heuristics from Intention Models -- Extended Version

Authors: Giacomo Rosa, Jean Honorio, Nir Lipovetzky, Sebastian Sardina

Organizations: The University of Melbourne · RMIT University

Abstract

Classical planning aims to find a sequence of actions, a plan, that maps a starting state into one of the goal states. If a trajectory appears to be leading to the goal, should we prioritise exploring it? Seminal work in goal recognition (GR) has defined GR in terms of a classical planning problem, adopting classical solvers and heuristics to recognise plans. We come full circle, and study the adoption and properties of GR-derived heuristics for seeking solutions to classical planning problems. We propose a new divergence-based framework for assessing goal intention, which informs a new class of efficiently-computable heuristics. As a proof of concept, we derive two such heuristics, and show that they can already yield improvements for top-scoring classical planners. Our work provides foundational knowledge for understanding and deriving probabilistic intention-based heuristics for planning.

Figures & tables

	$f5$	$f5$ -ol ${}^{rp}_{r}$	$f5$	$f5$ -op ${}^{rp}_{r}$	$f5$ -ol ${}^{rp}_{r}$	$f5$ -op ${}^{rp}_{r}$
Coverage (1831)	1,510	1,560 (5.03)	1,510	1,556 (1.5)	1,560 (5.03)	1,556 (1.5)
% Score	76.77%	80.20% (0.35)	76.77%	79.90% (0.15)	80.20% (0.35)	79.90% (0.15)
N Fewer Expansions	368.2 (6.7)	1,134.4 (8.7)	341.4 (9.1)	1,159.4 (8.0)	412.0 (7.3)	716.6 (6.5)
N Lower Time	582.4 (99.5)	897.0 (92.2)	658.8 (117.3)	814.6 (102.4)	811.8 (130.1)	658.0 (122.6)
N Lower Plan Cost	760.6 (8.9)	369.6 (7.6)	763.6 (2.9)	375.8 (1.7)	375.2 (4.5)	349.2 (13.4)
Average EpS	40,111	35,680	40,111	33,626	35,680	33,626

Table 1: Enhancing a basic BFWS(

f5

) solver with proposed ol

{}^{rp}_{r}

and op

{}^{rp}_{r}

heuristics. % score is the average of the % of instances solved in each problem domain. Results for comparisons and solvers with a randomised component represent the mean, and include the standard deviation across 5 measurements. N Fewer Expansions, N Lower Time and N Lower Plan Cost represent the number of problems where a variant scores better in the respective metrics (ignoring ties). Avg. EpS represents the average number of expansions per second across problems solved by all planners, and proxies computational overhead. Results indicate that ol

{}^{rp}_{r}

and op

{}^{rp}_{r}

reduce the number of expansions across a significant portion of problems, on average improving also coverage and solution time. Such improvements come at the cost of on average worse plan quality compared to the baseline.

Planner	Coverage (1831)	% score	Agile score
Dual-BFWS	1,607	83.6%	1,200.8
ApxNoveltyT	1,611 (3.5)	83.8% (0.2)	1,233.7 (0.2)
LAMA	1,535	79.1%	1,192.3
Scorpion-M	1,591	82.9%	1,206.4
OL ${}_{r}^{rp}$	1,621 (3.2)	84.6% (0.3)	1,229.4 (3.4)
OL ${}_{r}^{rp}$ -UTP	1,616 (2.1)	84.2% (0.1)	1,236.0 (1.9)

Table 2: Mean coverage and Agile score of enhanced variants. Our enhancements of BFWS(

f5)

outperform SoTA planners, on average solving more problems, more quickly, without using multiple open lists or runs. OL

{}_{r}^{rp}

-Dual provides an improved dual-strategy solver, solving more problems than all other tested planners and attaining a meaningfully higher Agile score than BFNoS-Dual, the second-best in coverage.

Panel of Figure 1 — Figure 1: Number of expansions (x-axis vs. y-axis) (lower is better). Comparing BFWS( $f5$ ), BFWS( $f5$ )-ol ${}_{r}^{rp}$ , and BFWS( $f5$ )-op ${}_{r}^{rp}$ . The high-density regions lie below the red line, indicating a systematic improvement over the baseline BFWS( $f5$ ).

Panel of Figure 2 — Figure 2: Total solver time (x-axis vs. y-axis) (lower is better). Comparing BFWS( $f5$ ), BFWS( $f5$ )-ol ${}_{r}^{rp}$ , and BFWS( $f5$ )-op ${}_{r}^{rp}$ .

Panel of Figure 3 — Figure 3: Plan cost (Unit action costs) (x-axis vs. y-axis) (lower is better). Comparing BFWS( $f5$ ), BFWS( $f5$ )-ol ${}_{r}^{rp}$ , and BFWS( $f5$ )-op ${}_{r}^{rp}$ .

Domain	BFWS( $f5$ )	BFWS( $f5$ )-ol ${}_{r}^{rp}$	BFWS( $f5$ )-op ${}_{r}^{rp}$
agricola-sat18-strips (20)	10	12 $\pm$ 0.6	12 $\pm$ 0.8
airport (50)	47	47 $\pm$ 0.6	48 $\pm$ 0.0
assembly (30)	30	30 $\pm$ 0.5	29 $\pm$ 0.5
barman-sat14-strips (20)	20	20 $\pm$ 0.0	20 $\pm$ 0.0
blocks (35)	35	35 $\pm$ 0.0	35 $\pm$ 0.0
caldera-sat18-adl (20)	15	18 $\pm$ 0.6	16 $\pm$ 1.0

Table 3: Comparative performance analysis across the full set of benchmark domains. % score is the average of the % of instances solved in each domain. Values for solvers with randomized components represent the mean and include the standard deviation across 5 measurements, using seeds from 0 to 4.

Domain	Dual-	Apx-BFWS	LAMA-	Scorpion	BFNoS-Dual	BFWS t -hlm	BFWS t -hlm	BFWS t -hlm	BFWS t -hlm
	BFWS	(Tarski)	First	Maidu		op ${}_{r}^{rp}$	ol ${}_{r}^{rp}$	ol ${}_{r}^{rp}$ -UTP	ol ${}_{r}^{rp}$ -Dual
agricola-sat18-strips	13	18 $\pm$ 0.6	12	12	15 $\pm$ 0.0	13 $\pm$ 1.5	14 $\pm$ 1.0	14 $\pm$ 0.6	14 $\pm$ 0.6
airport	46	47 $\pm$ 0.6	34	38	46 $\pm$ 0.6	47 $\pm$ 0.0	47 $\pm$ 0.0	47 $\pm$ 0.0	47 $\pm$ 0.0
assembly	30	30 $\pm$ 0.0	30	30	30 $\pm$ 0.0	30 $\pm$ 0.0	30 $\pm$ 0.0	30 $\pm$ 0.6	30 $\pm$ 0.0
barman-sat14-strips	20	20 $\pm$ 0.0	20	20	20 $\pm$ 0.0	20 $\pm$ 0.0	20 $\pm$ 0.0	20 $\pm$ 0.0	20 $\pm$ 0.0
blocks	35	35 $\pm$ 0.0	35	35	35 $\pm$ 0.0	35 $\pm$ 0.0	35 $\pm$ 0.0	35 $\pm$ 0.0	35 $\pm$ 0.0

Table 4: Comparative performance analysis across the full set of benchmark domains. % score is the average of the % of instances solved in each domain. Values for solvers with randomized components represent the mean and include the standard deviation across 3 measurements, using seeds from 0 to 2.

Symbolic Planning

Symbolic planning represents goals, states, and actions explicitly, then searches for action sequences that satisfy constraints.

See all 89 papers

Removing Planner Bias in Goal Recognition Through Multi-Plan Dataset Generation
Feb 16, 2026Mustafa F. Abdelwahed, Felipe Meneguzzi Kin Max Piamolini Gusmao, Joan Espasa
University of St Andrews, School of Computer Science, UK · Pontifical Catholic University of Rio Grande do Sul, Brazil · University of Aberdeen, UK
A Probabilistic Framework for Hierarchical Goal Recognition
Apr 24, 2026Chenyuan Zhang, Katherine Ip, Hamid Rezatofighi +2
Monash University · The University of Melbourne
A complementary study on PlanGPT: Evaluation with defined Performance Metrics and comparison with a planner
Jun 9, 2026Youssef Abdelkader, Humbert Fiorino, Damien PellierLLM Planning
Univ. Grenoble Alpes - LIG Grenoble, France

Feb 16, 2026J/K move · Enter open · S save

Removing Planner Bias in Goal Recognition Through Multi-Plan Dataset Generation

Mustafa F. Abdelwahed, Felipe Meneguzzi Kin Max Piamolini Gusmao, Joan Espasa

University of St Andrews, School of Computer Science, UK · Pontifical Catholic University of Rio Grande do Sul, Brazil · University of Aberdeen, UK

Open paper AskPDF arXiv

Autonomous agents require some form of goal and plan recognition to interact in multiagent settings. Unfortunately, all existing goal recognition datasets suffer from a systematical bias induced by the planning systems that generated them, namely heuristic-based forward search. This means that existing datasets lack enough challenge for more realistic scenarios (e.g., agents using different planners), which impacts the evaluation of goal recognisers with respect to using different planners for the same goal. In this paper, we propose a new method that uses top-k planning to generate multiple, different, plans for the same goal hypothesis, yielding benchmarks that mitigate the bias found in the current dataset. This allows us to introduce a new metric called Version Coverage Score (VCS) to measure the resilience of the goal recogniser when inferring a goal based on different sets of plans. Our results show that the resilience of the current state-of-the-art goal recogniser degrades substantially under low observability settings.

Abstract

Symbolic Planning

Explore similar work

Removing Planner Bias in Goal Recognition Through Multi-Plan Dataset Generation

A Probabilistic Framework for Hierarchical Goal Recognition

A complementary study on PlanGPT: Evaluation with defined Performance Metrics and comparison with a planner