Resolving Over-Constrained Probabilistic Temporal Problems through Chance Constraint Relaxation
Authors: Peng Yu, Cheng Fang, Brian Williams
AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In the rest of the section, we present experimental results that demonstrate the run-time performance of CDCR on problems with different size and complexity. The test cases were generated using a mission simulator for underwater expeditions. We benchmarked CDCR on each problem twice: the first run assumes normal distribution over the uncertain durations, and the second run assumes uniform distribution. In each test, the time consumption of the first solution returned was recorded. The results are shown in Figure 6. |
| Researcher Affiliation | Academia | Peng Yu and Cheng Fang and Brian Williams Massachusetts Institute of Technology Computer Science and Artificial Intelligence Laboratory 32 Vassar Street, Cambridge, MA 02139 {yupeng,cfang,williams}@mit.edu |
| Pseudocode | Yes | Algorithm 1: The CDCR algorithm; Algorithm 2: Addition to CDCR for Incorporating Inputs |
| Open Source Code | No | The paper does not provide any links to open-source code for the described methodology or state that code is made available. |
| Open Datasets | No | The paper states: "The test cases were generated using a mission simulator for underwater expeditions." It does not provide concrete access information (link, DOI, specific repository, or formal citation for a public dataset) for the generated test cases. |
| Dataset Splits | No | The paper mentions running CDCR on "2000 test cases" but does not specify any training, validation, or test dataset splits, nor does it refer to cross-validation. |
| Hardware Specification | No | The paper mentions that CDCR resolved most problems with less than 100 constraints within 30 seconds and discusses computational time, but it does not provide any specific hardware details like CPU/GPU models or memory specifications used for the experiments. |
| Software Dependencies | No | The paper mentions using "the Interior Point Optimizer (IPOPT (Wchter and Biegler 2006)) to compute optimal resolutions" but does not provide a version number for IPOPT or any other software dependencies. |
| Experiment Setup | Yes | The test cases were generated using a mission simulator for underwater expeditions. Given a set of target locations on a map, the simulator generates survey tasks around them and connects these locations with transit activities. Each test case describes a dive of multiple survey tasks: the traversals are represented by probabilistic durations, while the survey times and battery restrictions are modeled by simple temporal constraints. The operational risk limit is specified by the chance constraints in the cc-p STPs. In total, we created 2000 test cases using randomly generated numbers of vehicles, risk bounds, task locations and mission length. For each problem, we run CDCR to find a grounded STNU of the cc-p STP that is dynamically controllable while meeting the chance constraint, or a set of relaxations for the cc-p STP that will enable such a STNU. The timeout for each test is 30 seconds, which is usually the maximum duration that users are willing to wait for. The cutoff for the uniform distribution is selected using µ 2σ. |