Heuristic Search for Multi-Objective Probabilistic Planning

Authors: Dillon Z. Chen, Felipe Trevizan, Sylvie Thiébaux

AAAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 5 Experiments In this section we empirically evaluate the different algorithms and heuristics for MOSSPs in several domains.
Researcher Affiliation Academia Dillon Z. Chen1, Felipe Trevizan1, Sylvie Thi ebaux1,2 1School of Computing, The Australian National University 2LAAS-CNRS, ANITI, Universit e de Toulouse {Dillon.Chen, Felipe.Trevizan, Sylvie.Thiebaux}@anu.edu.au
Pseudocode Yes Algorithm 1: MOVI, Algorithm 2: MOVI under Assumption 1, Algorithm 3: MOLAO, Algorithm 4: IMOLAO, Algorithm 5: MOLRTDP
Open Source Code Yes We implemented the MO versions of the VI, TVI, (i)LAO and LRTDP algorithms and the MO version of the PDB abstraction heuristics (Hpdb mossp) in C++.2 PDB heuristics are computed using TVI, ε = 0.001 and b = 100.2 Code at https://github.com/Dillon ZChen/cpp-mossp-planner
Open Datasets No Since no benchmark domains for MOSSPs exist, we adapt domains from a variety of sources to capture challenging features of both SSPs and MO deterministic planning. (No direct public access information for the adapted datasets is provided.)
Dataset Splits No No specific dataset split information (like percentages or sample counts) for training, validation, or testing is provided. The paper discusses problem domains rather than traditional datasets with splits.
Hardware Specification Yes The experiments are run on a cluster with Intel Xeon 3.2 GHz CPUs.
Software Dependencies Yes We used CPLEX version 22.1 as the LP solver for computing CCS.
Experiment Setup Yes The consistency threshold is set to ε = 0.001 and we set b = 100.