Heavy-Head Sampling for Fast Imitation Learning of Machine Learning Based Combinatorial Auction Solver

Chen, Peng; Liao, Bolin

doi:10.1007/s11063-022-10900-y

Cited by 11 publications

(11 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed weighted column sampling (WCS) optimization framework can be divided into creating, solving, and merging the solutions of multiple smaller IPs, which are referred to as the WCS models. Each WCS model 𝑖 is derived from the MS model ( 4)-( 8) by considering only a subset N 𝑖 ⊆ N of observations, and using only a subset P 𝑒 𝑖 ⊆ P 𝑒 of feature variables 𝑏 𝑝 that are sampled according to their SIS values (17). The overall procedure of the WCS optimization is shown in Figure 1, and the major steps are described as follows.…”

Section: Weighted Column Sampling Optimizationmentioning

confidence: 99%

Explaining with Greater Support: Weighted Column Sampling Optimization for q-Consistent Summary-Explanations

Chen¹,

Dai²,

Xia³

et al. 2023

Preprint

View full text Add to dashboard Cite

Machine learning systems have been extensively used as auxiliary tools in domains that require critical decision-making, such as healthcare and criminal justice. The explainability of decisions is crucial for users to develop trust on these systems. In recent years, the globally-consistent rule-based summary-explanation and its max-support (MS) problem have been proposed, which can provide explanations for particular decisions along with useful statistics of the dataset. However, globally-consistent summary-explanations with limited complexity typically have small supports, if there are any. In this paper, we propose a relaxed version of summaryexplanation, i.e., the 𝑞-consistent summary-explanation, which aims to achieve greater support at the cost of slightly lower consistency.The challenge is that the max-support problem of 𝑞-consistent summary-explanation (MSqC) is much more complex than the original MS problem, resulting in over-extended solution time using standard branch-and-bound solvers. To improve the solution time efficiency, this paper proposes the weighted column sampling (WCS) method based on solving smaller problems by sampling variables according to their simplified increase support (SIS) values. Experiments verify that solving MSqC with the proposed SIS-based WCS method is not only more scalable in efficiency, but also yields solutions with greater support and better global extrapolation effectiveness.

show abstract

Section: Weighted Column Sampling Optimizationmentioning

confidence: 99%

Explaining with Greater Support: Weighted Column Sampling Optimization for q-Consistent Summary-Explanations

Chen¹,

Dai²,

Xia³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…This representation is natural for MILPs and has shown promising performance. In [16], Peng et al proposed that prioritizing the sampling of certain branching decisions over others and thus providing a better branching data distribution could further improve the performance of the trained model. In [17], the authors pointed out that the GCNN-based approach relies too heavily on high-end GPU, which may not be be accessible to many practitioners.…”

Section: The Bigraph Representation For State Embeddingmentioning

confidence: 99%

“…As in [16], the results in this paper are presented in the form of "mean r ± std%" to avoid the dependence of results on different experimental environments, and "r" is the mean of Node or Time as a reference value. For example, 0.7883r ± 6.68% means that the metric is 0.7883 times the reference value, and the per-instance standard deviation is 0.0668 averaged over all instances.…”

Section: Comparison Of Problem-solving Efficiencymentioning

confidence: 99%

“…In [14], a variable-constraint bipartite graph (or bigraph) representation was leveraged for B&B state embedding, and a graph convolutional neural network (GCNN) model was proposed for learning the branching policy. The bigraph representation is natural for MILPs and has shown promising performance [16,17]. However, the bigraph representation was designed for general MILP problems, i.e., aiming to apply one ML model to many different types of MILP problems.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Batch-Wise Permutation Feature Importance Evaluation and Problem-Specific Bigraph for Learn-to-Branch

2022

Self Cite

View full text Add to dashboard Cite

The branch-and-bound algorithm for combinatorial optimization typically relies on a plethora of handcraft expert heuristics, and a research direction, so-called learn-to-branch, proposes to replace the expert heuristics in branch-and-bound with machine learning models. Current studies in this area typically use an imitation learning (IL) approach; however, in practice, IL often suffers from limited training samples. Thus, it has been emphasized that a small-dataset fast-training scheme for IL in learn-to-branch is worth studying, so that other methods, e.g., reinforcement learning, may be used for subsequent training. Thus, this paper focuses on the IL part of a mixed training approach, where a small-dataset fast-training scheme is considered. The contributions are as follows. First, to compute feature importance metrics so that the state-of-the-art bigraph representation can be effectively reduced for each problem type, a batch-wise permutation feature importance evaluation method is proposed, which permutes features within each batch in the forward pass. Second, based on the evaluated importance of the bigraph features, a reduced bigraph representation is proposed for each of the benchmark problems. The experimental results on four MILP benchmark problems show that our method improves branching accuracy by 8% and reduces solution time by 18% on average under the small-dataset fast-training scheme compared to the state-of-the-art bigraph-based learn-to-branch method. The source code is available online at GitHub.

show abstract

“…The authors in [40] proposed a hybrid Ant Colony Optimization (ACO) algorithm to solve the NP-hard nature combinatorial auction problem at the expense of more execution time. In [41], a heavy-head sampling strategy was proposed with Imitation Learning (IL) to solve the WDP of combinatorial auction. They proposed IL used with Reinforcement Learning (RL) to improve the evaluation process of CA; however, it is prone to extended training time.…”

Section: B Literature Reviewmentioning

confidence: 99%

A Computationally Efficient Method for Energy Allocation in Spot Markets With Application to Transactive Energy Systems

et al. 2022

View full text Add to dashboard Cite

Spot markets provide an interesting opportunity for profit maximization by energy trading based on immediate decisions on participant bids. However, their short market-clearing time can affect computational efficiency, search space, and reliability of price-energy allocation to bidding participants. Accordingly, developing a prompt and effective decision-making process plays a vital role in smooth energy delivery in these markets. This paper proposes an approach to alleviate the computational cost of the spot market aggregator in order to decide price-energy bids. The proposed bidding model is developed for the transactive energy systems, where the spot market aggregator utilizes the proposed method to maximize profit by choosing participants' demand-side bids. The proposed method can efficiently manage participants' combined energy and price information and avoid a highly complicated search space. It takes advantage of the multi-variable Taylor series approximation to create users' individual cost functions. The approximated cost functions lead to user-specific bids that expedite the spot market transaction while maintaining aggregator profit. The resultant system is able to exercise profit maximization with high performance within milliseconds. The efficiency of this scheme is also demonstrated through a comparative study by using the particle swarm optimization method.INDEX TERMS Transactive energy, spot markets, combinatorial auction, smart grids, computational efficiency, demand response. NOMENCLATURE INDICES FUNCTIONS xi Average energy demand. C(y) Original cost function. C aprx Approximation of cost function. U i k+1 Utility function.

show abstract

Heavy-Head Sampling for Fast Imitation Learning of Machine Learning Based Combinatorial Auction Solver

Cited by 11 publications

References 30 publications

Explaining with Greater Support: Weighted Column Sampling Optimization for q-Consistent Summary-Explanations

Explaining with Greater Support: Weighted Column Sampling Optimization for q-Consistent Summary-Explanations

Batch-Wise Permutation Feature Importance Evaluation and Problem-Specific Bigraph for Learn-to-Branch

A Computationally Efficient Method for Energy Allocation in Spot Markets With Application to Transactive Energy Systems

Contact Info

Product

Resources

About