Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1]
Fast Learning for Renewal Optimization in Online Task Scheduling
Authors: Michael J. Neely
JMLR 2021 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | This section presents simulations of the proposed algorithm under the initial condition θ[0] = θmin and stepsize η[k] = 1 (k+2)Tmin . Fig. 3 compares the proposed algorithm with the greedy strategy for two different values of p. Data is averaged over 5000 independent sample paths. |
| Researcher Affiliation | Academia | Michael J. Neely EMAIL Department of Electrical Engineering University of Southern California Los Angeles, CA, 90089-2565, USA |
| Pseudocode | No | The paper describes the iterative algorithm in Section 4 using numbered steps and mathematical equations (24)-(26). It details the logic as: 'On each frame k {0, 1, 2, . . .} do: Observe S[k] ΩS and the current θ[k] value. Choose (T[k], R[k]) to solve: Maximize: R[k] θ[k]T[k] (24) Subject to: (T[k], R[k]) D(S[k]) (25) breaking ties arbitrarily. Update θ[k] via the iteration: θ[k + 1] = [θ[k] + η[k](R[k] θ[k]T[k])]θmax θmin (26)'. This is a textual description with equations, not a structured pseudocode or algorithm block. |
| Open Source Code | No | The paper does not contain any explicit statements or links indicating that the source code for the described methodology is publicly available. |
| Open Datasets | No | The paper describes hypothetical systems for simulation, generating data based on specified distributions and parameters, rather than utilizing or providing access to pre-existing public datasets. For instance, in Section 8.1, it states: 'On each frame k we receive N[k] new potential projects, where N[k] {0, 1, 2, 3} with P[N[k] = i] = pi and p0 = 0.1, p1 = 0.9 p, p2 = p/2, p3 = p/2 where p [0, 0.9] is a parameter varied in the simulations... The vectors (Tj, Rj) for j {1, . . . , i} are generated independently with Tj Uniform([1, 10]) and Rj = Aj Tj where Aj Unif([0, 50]).' |
| Dataset Splits | No | The paper uses synthetically generated data for simulations, with '5000 independent sample paths' being averaged. It does not use pre-collected datasets that would require explicit training/validation/test splits. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used to run the simulations or experiments. |
| Software Dependencies | No | The paper does not provide specific software names with version numbers or library dependencies used for implementing the algorithms and running simulations. |
| Experiment Setup | Yes | The simulation section (Section 8) provides specific details about the experimental setup for each system. For System 1, it states: 'The proposed algorithm uses [θmin, θmax] = [0, 50] and Tmin = 1.' For System 2: 'The proposed algorithm uses [θmin, θmax] = [1, 2] and Tmin = 1.' For System 3: 'We use [θmin, θmax] = [1, 3], Tmin = 1.' It also specifies the initial condition as 'initial condition θ[0] = θmin and stepsize η[k] = 1 (k+2)Tmin .' |