Pub Date : 2023-07-01DOI: 10.1609/icaps.v33i1.27210
Francesco Percassi, Enrico Scala, M. Vallati
The plan, execution, and replan framework has proven to be extremely valuable in complex real-world applications, where the dynamics of the environment cannot be fully encoded in the domain model. However, this comes at the cost of regenerating plans from scratch, which can be expensive when expressive formalisms like PDDL+ are used. Given the complexity of generating PDDL+ plans, it would be ideal to reuse as much as possible of an existing plan, rather than generating a new one from scratch every time. To support more effective exploitation of the plan, execution, and replan framework in PDDL+, in this paper, we introduce the problem of discretized PDDL+ plan fixing, which allows one to fix existing plans according to some defined constraints. We demonstrate the theoretical implications of the introduced notion and introduce reformulations to address the problem using domain-independent planning engines. Our results show that such reformulations can outperform replanning from scratch and unlock planning engines to solve more problems with fine-grained discretizations.
{"title":"Fixing Plans for PDDL+ Problems: Theoretical and Practical Implications","authors":"Francesco Percassi, Enrico Scala, M. Vallati","doi":"10.1609/icaps.v33i1.27210","DOIUrl":"https://doi.org/10.1609/icaps.v33i1.27210","url":null,"abstract":"The plan, execution, and replan framework has proven to be extremely valuable in complex real-world applications, where the dynamics of the environment cannot be fully encoded in the domain model. However, this comes at the cost of regenerating plans from scratch, which can be expensive when expressive formalisms like PDDL+ are used. Given the complexity of generating PDDL+ plans, it would be ideal to reuse as much as possible of an existing plan, rather than generating a new one from scratch every time. To support more effective exploitation of the plan, execution, and replan framework in PDDL+, in this paper, we introduce the problem of discretized PDDL+ plan fixing, which allows one to fix existing plans according to some defined constraints. We demonstrate the theoretical implications of the introduced notion and introduce reformulations to address the problem using domain-independent planning engines. Our results show that such reformulations can outperform replanning from scratch and unlock planning engines to solve more problems with fine-grained discretizations.","PeriodicalId":239898,"journal":{"name":"International Conference on Automated Planning and Scheduling","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126625821","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1609/icaps.v33i1.27180
Clemens Büchner, Thomas Keller, Salomé Eriksson, M. Helmert
The computation of high-quality landmarks and orderings for heuristic state-space search is often prohibitively expensive to be performed in every generated state. Computing information only for the initial state and progressing it from every state to its successors is a successful alternative, exploited for example in classical planning by the LAMA planner. We propose a general framework for using landmarks in any kind of best-first search. Its core component, the progression function, uses orderings and search history to determine which landmarks must still be achieved. We show that the progression function that is used in LAMA infers invalid information in the presence of reasonable orderings. We define a sound progression function that allows to exploit reasonable orderings in cost-optimal planning and show empirically that our new progression function is beneficial both in satisficing and optimal planning.
{"title":"Landmark Progression in Heuristic Search","authors":"Clemens Büchner, Thomas Keller, Salomé Eriksson, M. Helmert","doi":"10.1609/icaps.v33i1.27180","DOIUrl":"https://doi.org/10.1609/icaps.v33i1.27180","url":null,"abstract":"The computation of high-quality landmarks and orderings for heuristic\u0000state-space search is often prohibitively expensive to be performed in\u0000every generated state. Computing information only for the initial\u0000state and progressing it from every state to its successors is a\u0000successful alternative, exploited for example in classical planning by\u0000the LAMA planner. We propose a general framework for using landmarks\u0000in any kind of best-first search. Its core component, the progression\u0000function, uses orderings and search history to determine which\u0000landmarks must still be achieved. We show that the progression\u0000function that is used in LAMA infers invalid information in the\u0000presence of reasonable orderings. We define a sound progression\u0000function that allows to exploit reasonable orderings in cost-optimal\u0000planning and show empirically that our new progression function is\u0000beneficial both in satisficing and optimal planning.","PeriodicalId":239898,"journal":{"name":"International Conference on Automated Planning and Scheduling","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128975246","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1609/icaps.v33i1.27191
Dennis Gross, C. Schmidl, N. Jansen, G. Pérez
Cooperative multi-agent reinforcement learning (CMARL) enables agents to achieve a common objective. However, the safety (a.k.a. robustness) of the CMARL agents operating in critical environments is not guaranteed. In particular, agents are susceptible to adversarial noise in their observations that can mislead their decision-making. So-called denoisers aim to remove adversarial noise from observations, yet, they are often error-prone. A key challenge for any rigorous safety verification technique in CMARL settings is the large number of states and transitions, which generally prohibits the construction of a (monolithic) model of the whole system. In this paper, we present a verification method for CMARL agents in settings with or without adversarial attacks or denoisers. Our method relies on a tight integration of CMARL and a verification technique referred to as model checking. We showcase the applicability of our method on various benchmarks from different domains. Our experiments show that our method is indeed suited to verify CMARL agents and that it scales better than a naive approach to model checking.
{"title":"Model Checking for Adversarial Multi-Agent Reinforcement Learning with Reactive Defense Methods","authors":"Dennis Gross, C. Schmidl, N. Jansen, G. Pérez","doi":"10.1609/icaps.v33i1.27191","DOIUrl":"https://doi.org/10.1609/icaps.v33i1.27191","url":null,"abstract":"Cooperative multi-agent reinforcement learning (CMARL) enables agents to achieve a common objective. \u0000However, the safety (a.k.a. robustness) of the CMARL agents operating in critical environments is not guaranteed. \u0000In particular, agents are susceptible to adversarial noise in their observations that can mislead their decision-making.\u0000So-called denoisers aim to remove adversarial noise from observations, yet, they are often error-prone.\u0000A key challenge for any rigorous safety verification technique in CMARL settings is the large number of states and transitions, which generally prohibits the construction of a (monolithic) model of the whole system.\u0000In this paper, we present a verification method for CMARL agents in settings with or without adversarial attacks or denoisers.\u0000Our method relies on a tight integration of CMARL and a verification technique referred to as model checking.\u0000We showcase the applicability of our method on various benchmarks from different domains.\u0000Our experiments show that our method is indeed suited to verify CMARL agents and that it scales better than a naive approach to model checking.","PeriodicalId":239898,"journal":{"name":"International Conference on Automated Planning and Scheduling","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130147427","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1609/icaps.v33i1.27212
Quang Anh Pham, H. Lau, Minh Hoàng Hà, Lam Vu
The traveling salesman problem (TSP) is the most well-known problem in combinatorial optimization which has been studied for many decades. This paper focuses on dealing with one of the most difficult TSP variants named the quadratic traveling salesman problem (QTSP) that has numerous planning applications in robotics and bioinformatics. The goal of QTSP is similar to TSP which finds a cycle visiting all nodes exactly once with minimum total costs. However, the costs in QTSP are associated with three vertices traversed in succession (instead of two like in TSP). This leads to a quadratic objective function that is much harder to solve. To efficiently solve the problem, we propose a hybrid genetic algorithm including a local search procedure for intensification and a new mutation operator for diversification. The local search is composed of a restricted double-bridge move (a variant of 4-Opt); and we show the neighborhood can be evaluated in O(n^2), the same complexity as for the classical TSP. The mutation phase is inspired by a ruin-and-recreate scheme. Experimental results conducted on benchmark instances show that our method significantly outperforms state-of-the-art algorithms in terms of solution quality. Out of 800 considered instances, it finds 437 new best-known solutions.
{"title":"An Efficient Hybrid Genetic Algorithm for the Quadratic Traveling Salesman Problem","authors":"Quang Anh Pham, H. Lau, Minh Hoàng Hà, Lam Vu","doi":"10.1609/icaps.v33i1.27212","DOIUrl":"https://doi.org/10.1609/icaps.v33i1.27212","url":null,"abstract":"The traveling salesman problem (TSP) is the most well-known problem in combinatorial optimization which has been studied for many decades. This paper focuses on dealing with one of the most difficult TSP variants named the quadratic traveling salesman problem (QTSP) that has numerous planning applications in robotics and bioinformatics. The goal of QTSP is similar to TSP which finds a cycle visiting all nodes exactly once with minimum total costs. However, the costs in QTSP are associated with three vertices traversed in succession (instead of two like in TSP). This leads to a quadratic objective function that is much harder to solve. To efficiently solve the problem, we propose a hybrid genetic algorithm including a local search procedure for intensification and a new mutation operator for diversification. The local search is composed of a restricted double-bridge move (a variant of 4-Opt); and we show the neighborhood can be evaluated in O(n^2), the same complexity as for the classical TSP. The mutation phase is inspired by a ruin-and-recreate scheme. Experimental results conducted on benchmark instances show that our method significantly outperforms state-of-the-art algorithms in terms of solution quality. Out of 800 considered instances, it finds 437 new best-known solutions.","PeriodicalId":239898,"journal":{"name":"International Conference on Automated Planning and Scheduling","volume":"109 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121997000","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1609/icaps.v33i1.27193
Guanghua Hu, Tim Miller, N. Lipovetzky
Epistemic planning plays an important role in multi-agent and human-agent interaction domains. Most existing works solve multi-agent epistemic planning problems by either pre-compiling them into classical planning problems; or, using explicit actions and their effects to encode Kripke-based semantics. A recent approach called Planning with Perspectives (PWP) delegates epistemic reasoning in planning to external functions using F-STRIPS, keeping the search within the planning algorithm and lazily evaluating epistemic formulae. Although PWP is expressive and efficient, it models S5 epistemic logic and does not support belief, including false belief. In this paper, we extend the PWP model to handle multi-agent belief by following the intuition that agents believe something they have seen until they see otherwise. We call this justified perspectives. We formalise this notion of multi-agent belief based on the definition of knowledge in PWP. Using experiments on existing epistemic and doxastic planning benchmarks, we show that our belief planner can solve benchmarks more efficiently than the state-of-the-art baseline, and can model some problems that are infeasible to model using propositional-based approaches.
{"title":"Planning with Multi-Agent Belief Using Justified Perspectives","authors":"Guanghua Hu, Tim Miller, N. Lipovetzky","doi":"10.1609/icaps.v33i1.27193","DOIUrl":"https://doi.org/10.1609/icaps.v33i1.27193","url":null,"abstract":"Epistemic planning plays an important role in multi-agent and human-agent interaction domains. \u0000Most existing works solve multi-agent epistemic planning problems by either pre-compiling them into classical planning problems; or, using explicit actions and their effects to encode Kripke-based semantics. \u0000A recent approach called Planning with Perspectives (PWP) delegates epistemic reasoning in planning to external functions using F-STRIPS, keeping the search within the planning algorithm and lazily evaluating epistemic formulae.\u0000\u0000Although PWP is expressive and efficient, it models S5 epistemic logic and does not support belief, including false belief. \u0000In this paper, we extend the PWP model to handle multi-agent belief by following the intuition that agents believe something they have seen until they see otherwise. We call this justified perspectives. We formalise this notion of multi-agent belief based on the definition of knowledge in PWP. Using experiments on existing epistemic and doxastic planning benchmarks, we show that our belief planner can solve benchmarks more efficiently than the state-of-the-art baseline, and can model some problems that are infeasible to model using propositional-based approaches.","PeriodicalId":239898,"journal":{"name":"International Conference on Automated Planning and Scheduling","volume":"56 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126140089","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1609/icaps.v33i1.27189
Daniel Gnad, M. Helmert, P. Jonsson, Alexander Shleyfman
Restricted Tasks (RT) are a special case of numeric planning characterized by numeric conditions that involve one numeric variable per formula and numeric effects that allow only the addition of constants. Despite this, RTs form an expressive class whose planning problem is undecidable. The restricted nature of RTs often makes problem modeling awkward and unnecessarily complicated. We show that this can be alleviated by compiling mathematical operations that are not natively supported into RTs using macro-like action sequences. With that, we can encode many features found in general numeric planning such as constant multiplication, addition of linear formulas, and integer division and residue. We demonstrate how our compilations can be used to capture challenging mathematical problems such as the (in)famous Collatz conjecture. Our approach additionally gives a simple undecidability proof for RTs, and the proof shows that the number of variables needed to construct an undecidable class of RTs is surprisingly low: two numeric and one propositional variable.
{"title":"Planning over Integers: Compilations and Undecidability","authors":"Daniel Gnad, M. Helmert, P. Jonsson, Alexander Shleyfman","doi":"10.1609/icaps.v33i1.27189","DOIUrl":"https://doi.org/10.1609/icaps.v33i1.27189","url":null,"abstract":"Restricted Tasks (RT) are a special case of numeric planning characterized by numeric conditions that involve one numeric variable per formula and numeric effects that allow only the addition of constants. Despite this, RTs form an expressive class whose planning problem is undecidable. The restricted nature of RTs often makes problem modeling awkward and unnecessarily complicated. We show that this can be alleviated by compiling mathematical operations that are not natively supported into RTs using macro-like action sequences. With that, we can encode many features found in general numeric planning such as constant multiplication, addition of linear formulas, and integer division and residue. We demonstrate how our compilations can be used to capture challenging mathematical problems such as the (in)famous Collatz conjecture. Our approach additionally gives a simple undecidability proof for RTs, and the proof shows that the number of variables needed to construct an undecidable class of RTs is\u0000surprisingly low: two numeric and one propositional variable.","PeriodicalId":239898,"journal":{"name":"International Conference on Automated Planning and Scheduling","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128513576","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The merge-and-shrink framework is a powerful tool to construct state space abstractions based on factored representations. One of its core applications in classical planning is the construction of admissible abstraction heuristics. In this paper, we develop a compositional theory of merge-and-shrink in the context of probabilistic planning, focusing on stochastic shortest path problems (SSPs). As the basis for this development, we contribute a novel factored state space model for SSPs. We show how general transformations, including abstractions, can be formulated on this model to derive admissible and/or perfect heuristics. To formalize the merge-and-shrink framework for SSPs, we transfer the fundamental merge-and-shrink transformations from the classical setting: shrinking, merging, and label reduction. We analyze the formal properties of these transformations in detail and show how the conditions under which shrinking and label reduction lead to perfect heuristics can be extended to the SSP setting.
{"title":"A Theory of Merge-and-Shrink for Stochastic Shortest Path Problems","authors":"Thorsten Klößner, Á. Torralba, Marcel Steinmetz, Silvan Sievers","doi":"10.1609/icaps.v33i1.27196","DOIUrl":"https://doi.org/10.1609/icaps.v33i1.27196","url":null,"abstract":"The merge-and-shrink framework is a powerful tool to construct state space abstractions based on factored representations. One of its core applications in classical planning is the construction of admissible abstraction heuristics. In this paper, we develop a compositional theory of merge-and-shrink in the context of probabilistic planning, focusing on stochastic shortest path problems (SSPs). As the basis for this development, we contribute a novel factored state space model for SSPs. We show how general transformations, including abstractions, can be formulated on this model to derive admissible and/or perfect heuristics. To formalize the merge-and-shrink framework for SSPs, we transfer the fundamental merge-and-shrink transformations from the classical setting: shrinking, merging, and label reduction. We analyze the formal properties of these transformations in detail and show how the conditions under which shrinking and label reduction lead to perfect heuristics can be extended to the SSP setting.","PeriodicalId":239898,"journal":{"name":"International Conference on Automated Planning and Scheduling","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132323375","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1609/icaps.v33i1.27236
Viet The Bui, Tien Mai
Recent works using deep reinforcement learning (RL) to solve routing problems such as the capacitated vehicle routing problem (CVRP) have focused on improvement learning-based methods, which involve improving a given solution until it becomes near-optimal. Although adequate solutions can be achieved for small problem instances, their efficiency degrades for large-scale ones. In this work, we propose a new improvement learning-based framework based on imitation learning where classical heuristics serve as experts to encourage the policy model to mimic and produce similar and better solutions. Moreover, to improve scalability, we propose Clockwise Clustering, a novel augmented framework for decomposing large-scale CVRP into subproblems by clustering sequentially nodes in clockwise order, and then learning to solve them simultaneously. Our approaches enhance state-of-the-art CVRP solvers while attaining competitive solution quality on several well-known datasets, including real-world instances with sizes up to 30,000 nodes. Our best methods are able to achieve new state-of-the-art solutions for several large instances and generalize to a wide range of CVRP variants and solvers. We also contribute new datasets and results to test the generalizability of our deep RL algorithms.
{"title":"Imitation Improvement Learning for Large-Scale Capacitated Vehicle Routing Problems","authors":"Viet The Bui, Tien Mai","doi":"10.1609/icaps.v33i1.27236","DOIUrl":"https://doi.org/10.1609/icaps.v33i1.27236","url":null,"abstract":"Recent works using deep reinforcement learning (RL) to solve routing problems such as the capacitated vehicle routing problem (CVRP) have focused on improvement learning-based methods, which involve improving a given solution until it becomes near-optimal. Although adequate solutions can be achieved for small problem instances, their efficiency degrades for large-scale ones. In this work, we propose a new improvement learning-based framework based on imitation learning where classical heuristics serve as experts to encourage the policy model to mimic and produce similar and better solutions. Moreover, to improve scalability, we propose Clockwise Clustering, a novel augmented framework for decomposing large-scale CVRP into subproblems by clustering sequentially nodes in clockwise order, and then learning to solve them simultaneously. Our approaches enhance state-of-the-art CVRP solvers while attaining competitive solution quality on several well-known datasets, including real-world instances with sizes up to 30,000 nodes. Our best methods are able to achieve new state-of-the-art solutions for several large instances and generalize to a wide range of CVRP variants and solvers. We also contribute new datasets and results to test the generalizability of our deep RL algorithms.","PeriodicalId":239898,"journal":{"name":"International Conference on Automated Planning and Scheduling","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123821522","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1609/icaps.v33i1.27233
Alberto Pozanco, Kassiani Papasotiriou, D. Borrajo, M. Veloso
Defining financial goals and formulating actionable plans to achieve them are essential components for ensuring financial health. This task is computationally challenging, given the abundance of factors that can influence one’s financial situation. In this paper, we present the Personal Finance Planner (PFP), which can generate personalized financial plans that consider a person’s context and the likelihood of taking financially related actions to help them achieve their goals. PFP solves the problem in two stages. First, it uses heuristic search to find a high-level sequence of actions that increase the income and reduce spending to help users achieve their financial goals. Next, it uses integer linear programming to determine the best low-level actions to implement the high-level plan. Results show that PFP is able to scale on generating realistic financial plans for complex tasks involving many low level actions and long planning horizons.
{"title":"Combining Heuristic Search and Linear Programming to Compute Realistic Financial Plans","authors":"Alberto Pozanco, Kassiani Papasotiriou, D. Borrajo, M. Veloso","doi":"10.1609/icaps.v33i1.27233","DOIUrl":"https://doi.org/10.1609/icaps.v33i1.27233","url":null,"abstract":"Defining financial goals and formulating actionable plans to achieve them are essential components for ensuring financial health. This task is computationally challenging, given the abundance of factors that can influence one’s financial situation. In this paper, we present the Personal Finance Planner (PFP), which can generate personalized financial plans that consider a person’s context and the likelihood of taking financially related actions to help them achieve their goals. PFP solves the problem in two stages. First, it uses heuristic search to find a high-level sequence of actions that increase the income and reduce spending to help users achieve their financial goals. Next, it uses integer linear programming to determine the best low-level actions to implement the high-level plan. Results show that PFP is able to scale on generating realistic financial plans for complex tasks involving many low level actions and long planning horizons.","PeriodicalId":239898,"journal":{"name":"International Conference on Automated Planning and Scheduling","volume":"202 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121796555","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1609/icaps.v33i1.27215
Philipp Sauer, Marcel Steinmetz, R. Künnemann, Jörg Hoffmann
In Stackelberg planning, a leader and a follower each choose a plan in the same planning task, the leader's objective being to maximize plan cost for the follower. This formulation naturally captures, among others, security-related scenarios where the leader defends an infrastructure against subsequent attacks by the follower. Indeed, Stackelberg planning has been applied to the analysis of email infrastructure security. At web scale, however, the planning tasks involved easily contain tens of thousands of objects, so that grounding becomes the bottleneck. Here we introduce a lifted form of Stackelberg planning to address this. We devise leader-follower search algorithms working at the level of the PDDL-style input model to the extent possible. Our experiments show that, in Stackelberg tasks with many objects, including in particular models of web infrastructure security, our lifted algorithms outperform grounded Stackelberg planning.
{"title":"Lifted Stackelberg Planning","authors":"Philipp Sauer, Marcel Steinmetz, R. Künnemann, Jörg Hoffmann","doi":"10.1609/icaps.v33i1.27215","DOIUrl":"https://doi.org/10.1609/icaps.v33i1.27215","url":null,"abstract":"In Stackelberg planning, a leader and a follower each choose a plan in the same planning task, the leader's objective being to maximize plan cost for the follower. This formulation naturally captures, among others, security-related scenarios where the leader defends an infrastructure against subsequent attacks by the follower. Indeed, Stackelberg planning has been applied to the analysis of email infrastructure security. At web scale, however, the planning tasks involved easily contain tens of thousands of objects, so that grounding becomes the bottleneck. Here we introduce a lifted form of Stackelberg planning to address this. We devise leader-follower search algorithms working at the level of the PDDL-style input model to the extent possible. Our experiments show that, in Stackelberg tasks with many objects, including in particular models of web infrastructure security, our lifted algorithms outperform grounded Stackelberg planning.","PeriodicalId":239898,"journal":{"name":"International Conference on Automated Planning and Scheduling","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134109346","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：