{"id":"https://openalex.org/W7203749867","doi":"https://doi.org/10.48550/arxiv.2608.17289","title":"PlanPO: Group Planning-Aware Policy Optimization for Multi-Turn Agentic LLMs","display_name":"PlanPO: Group Planning-Aware Policy Optimization for Multi-Turn Agentic LLMs","publication_year":2026,"publication_date":"2026-08-18","ids":{"openalex":"https://openalex.org/W7203749867","doi":"https://doi.org/10.48550/arxiv.2608.17289"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.17289","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.17289","pdf_url":"https://arxiv.org/pdf/2608.17289","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.17289","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5053996029","display_name":"Dayang Liang","orcid":"https://orcid.org/0000-0002-6181-6746","observed_orcids":["https://orcid.org/0000-0002-6181-6746"]},"institutions":[{"id":"https://openalex.org/I191208505","display_name":"Xiamen University","ror":"https://ror.org/00mcjh785","country_code":"CN","type":"education","lineage":["https://openalex.org/I191208505"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Liang, Dayang","raw_affiliation_strings":["Department of Automation , Xiamen University , Xiamen , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Automation , Xiamen University , Xiamen , China","institution_ids":["https://openalex.org/I191208505"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5147598835","display_name":"Liyuan He","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"He, Liyuan","raw_affiliation_strings":["School of Artificial Intelligence , Shanghai Jiaotong University , Shanghai , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Artificial Intelligence , Shanghai Jiaotong University , Shanghai , China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121612721","display_name":"Xuan Feng","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I159948400","display_name":"Jinan University","ror":"https://ror.org/02xe5ns62","country_code":"CN","type":"education","lineage":["https://openalex.org/I159948400"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Feng, Xuan","raw_affiliation_strings":["College of Cyberspace Security , Jinan University , Guangzhou , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"College of Cyberspace Security , Jinan University , Guangzhou , China","institution_ids":["https://openalex.org/I159948400"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125952054","display_name":"Shuxin Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Li, Shuxin","raw_affiliation_strings":["College of Computing and Data Science , Nanyang Technological University , Singapore"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"College of Computing and Data Science , Nanyang Technological University , Singapore","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123171041","display_name":"Bo An","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"An, Bo","raw_affiliation_strings":["College of Computing and Data Science , Nanyang Technological University , Singapore"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"College of Computing and Data Science , Nanyang Technological University , Singapore","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5126692004","display_name":"Yunlong Liu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I191208505","display_name":"Xiamen University","ror":"https://ror.org/00mcjh785","country_code":"CN","type":"education","lineage":["https://openalex.org/I191208505"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Liu, Yunlong","raw_affiliation_strings":["Department of Automation , Xiamen University , Xiamen , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Automation , Xiamen University , Xiamen , China","institution_ids":["https://openalex.org/I191208505"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.8579000234603882,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.8579000234603882,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10906","display_name":"AI-based Problem Solving and Planning","score":0.04800000041723251,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12031","display_name":"Speech and dialogue systems","score":0.027499999850988388,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/group-relative-policy-optimization","display_name":"Group Relative Policy Optimization","score":0.8240000009536743},{"id":"https://openalex.org/keywords/agentic-large-language-models","display_name":"agentic large language models","score":0.8360000252723694},{"id":"https://openalex.org/keywords/multi-turn-interaction","display_name":"multi-turn interaction","score":0.6710000038146973},{"id":"https://openalex.org/keywords/trajectory-length","display_name":"trajectory length","score":0.3659999966621399},{"id":"https://openalex.org/keywords/alfworld","display_name":"ALFWorld","score":0.6589999794960022},{"id":"https://openalex.org/keywords/webshop","display_name":"WebShop","score":0.9649999737739563}],"concepts":[{"id":"https://openalex.org/C148220186","wikidata":"https://www.wikidata.org/wiki/Q7111912","display_name":"Outcome (game theory)","level":2,"score":0.5181000232696533},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4948999881744385},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4359999895095825},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4027000069618225},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.37950000166893005},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.3723999857902527},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.36820000410079956},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.35420000553131104},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.34150001406669617},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.3301999866962433},{"id":"https://openalex.org/C2779436431","wikidata":"https://www.wikidata.org/wiki/Q30672407","display_name":"Policy learning","level":2,"score":0.3278000056743622},{"id":"https://openalex.org/C109986646","wikidata":"https://www.wikidata.org/wiki/Q546113","display_name":"Public policy","level":2,"score":0.30239999294281006},{"id":"https://openalex.org/C175444787","wikidata":"https://www.wikidata.org/wiki/Q39072","display_name":"Microeconomics","level":1,"score":0.3009999990463257},{"id":"https://openalex.org/C2781311116","wikidata":"https://www.wikidata.org/wiki/Q83306","display_name":"Group (periodic table)","level":2,"score":0.2994999885559082},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.29649999737739563},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.2946000099182129},{"id":"https://openalex.org/C100001284","wikidata":"https://www.wikidata.org/wiki/Q2248246","display_name":"Public economics","level":1,"score":0.2937000095844269},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.2849000096321106},{"id":"https://openalex.org/C2776035688","wikidata":"https://www.wikidata.org/wiki/Q1606558","display_name":"Affect (linguistics)","level":2,"score":0.2791999876499176},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.2768999934196472},{"id":"https://openalex.org/C79416737","wikidata":"https://www.wikidata.org/wiki/Q2305519","display_name":"Social learning","level":2,"score":0.2563999891281128}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.17289","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.17289","pdf_url":"https://arxiv.org/pdf/2608.17289","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.17289","doi":"https://doi.org/10.48550/arxiv.2608.17289","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.17289","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2608.17289","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.17289","pdf_url":"https://arxiv.org/pdf/2608.17289","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7203749867.pdf","grobid_xml":"https://content.openalex.org/works/W7203749867.grobid-xml"},"referenced_works_count":1,"referenced_works":["https://openalex.org/W7171552453"],"related_works":["https://openalex.org/W7113896486","https://openalex.org/W4417093501","https://openalex.org/W7140105616","https://openalex.org/W7207819884","https://openalex.org/W7120077454","https://openalex.org/W4414977868","https://openalex.org/W7165655450","https://openalex.org/W4416035922","https://openalex.org/W4414972137","https://openalex.org/W7160633711"],"abstract_inverted_index":{"Group-relative":[0],"policy":[1],"optimization":[2,115],"has":[3],"emerged":[4],"as":[5],"a":[6,69],"key":[7],"paradigm":[8],"for":[9,75,108],"training":[10,168],"agentic":[11],"large":[12],"language":[13],"models":[14],"(LLMs)":[15],"on":[16,104,149],"multi-turn":[17,154],"interactive":[18],"tasks.":[19],"However,":[20],"most":[21],"existing":[22],"variants":[23],"fail":[24],"to":[25,120],"distinguish":[26],"advantages":[27],"among":[28],"successful":[29,105],"trajectories":[30,34,106],"even":[31],"when":[32],"these":[33],"differ":[35],"substantially":[36],"in":[37,96],"their":[38],"interaction":[39,128],"efficiency.":[40],"For":[41],"instance,":[42],"circuitous":[43],"successes":[44],"are":[45],"often":[46],"assigned":[47],"the":[48,93,109,113,152],"identical":[49],"outcome":[50],"reward,":[51],"causing":[52],"advantage":[53,89],"collapse":[54],"and":[55,99,124,130,158],"severe":[56],"performance":[57],"bottlenecks.":[58],"To":[59],"this":[60,117],"end,":[61],"we":[62],"propose":[63],"Group":[64],"Planning-aware":[65],"Policy":[66],"Optimization":[67],"(PlanPO),":[68],"simple":[70],"yet":[71],"effective":[72],"RL":[73],"method":[74],"learning":[76],"generalizable":[77,123],"planning":[78,129],"abilities":[79],"beyond":[80],"task-specific":[81],"high-quality":[82,134],"behavior":[83],"patterns.":[84],"Specifically,":[85],"PlanPO":[86,143],"introduces":[87],"coarse-to-fine":[88],"signals,":[90],"which":[91],"capture":[92],"relative":[94],"differences":[95],"trajectory-level":[97],"lengths":[98,102],"turn-level":[100],"response":[101],"conditioned":[103],"sampled":[107],"same":[110],"task.":[111],"Within":[112],"group-relative":[114],"structure,":[116],"enables":[118],"agents":[119],"actively":[121],"learn":[122],"deliberate":[125],"behaviors":[126],"spanning":[127],"textual":[131],"generation":[132],"from":[133],"rollouts,":[135],"without":[136],"degenerating":[137],"into":[138],"vanilla":[139],"length":[140],"minimization.":[141],"Experimentally,":[142],"improves":[144],"over":[145],"GRPO":[146],"by":[147],"27.2\\%":[148],"average":[150],"across":[151],"challenging":[153],"benchmarks":[155],"ALFWorld,":[156],"WebShop,":[157],"SciWorld,":[159],"outperforming":[160],"recent":[161],"powerful":[162],"baselines":[163],"while":[164],"incurring":[165],"negligible":[166],"additional":[167],"cost.":[169]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-08-20T00:00:00"}