1,721,245 research outputs found
Relationship-Preserving E-Proceurement Auction
No full textWhile e-procurement auction has helped firms to achieve lower procurement costs, auction mechanisms that prevail at present in procurement markets need to address an important issue that concerns the ability to maintain long term relationships with the partners, especially in repeated e-procurement settings. In this paper, we propose a Relationship Preserving Auction (RPA) mechanism that augments the conventional auction mechanism with a bidder relationship scoring model. Our proposed mechanism gives increased chances of winning to the bidders who have bidden at relatively competitive price but had comparatively less wins so far. Keeping these bidders in the auction over time will lead to more competitive bidding prices and eventually reduce the auctioneer's total procurement cost in repeated auctions. From simulation experiments, we show how RPA works under different bidders' behavior. We show that RPA is able to obtain lower procurement cost compared to conventional procurement auctions when bidders bid opportunistically and renege readily to other markets
Multiobjective linear ensembles for robust and sparse training of few-bit neural networks
No full textTraining neural networks (NNs) using combinatorial optimization solvers has gained attention in recent years. In low-data settings, the use of state-of-the-art mixed integer linear programming solvers, for instance, has the potential to exactly train an NN while avoiding computing-intensive training and hyperparameter tuning and simultaneously training and sparsifying the network. We study the case of few-bit discrete-valued neural networks, both binarized neural networks (BNNs) whose values are restricted to +/- 1 and integer-valued neural networks (INNs) whose values lie in the range {-P,& mldr;,P}. Few-bit NNs receive increasing recognition because of their lightweight architecture and ability to run on low-power devices: for example, being implemented using Boolean operations. This paper proposes new methods to improve the training of BNNs and INNs. Our contribution is a multiobjective ensemble approach based on training a single NN for each possible pair of classes and applying a majority voting scheme to predict the final output. Our approach results in the training of robust sparsified networks whose output is not affected by small perturbations on the input and whose number of active weights is as small as possible. We empirically compare this BeMi approach with the current state of the art in solver-based NN training and with traditional gradient-based training, focusing on BNN learning in few-shot contexts. We compare the benefits and drawbacks of INNs versus BNNs, bringing new light to the distribution of weights over the {-P,& mldr;,P} interval. Finally, we compare multiobjective versus single-objective training of INNs, showing that robustness and network simplicity can be acquired simultaneously, thus obtaining better test performances. Although the previous state-of-the-art approaches achieve an average accuracy of 51.1% on the Modified National Institute of Standards and Technology data set, the BeMi ensemble approach achieves an average accuracy of 68.4% when trained with 10 images per class and 81.8% when trained with 40 images per class while having up to 75.3% NN links removed
Two-Echelon Repairable Item Inventory System with Limited Repair Capacity under Nonstationary Demands
Full text linkThe oil drilling model and iterative deepening genetic annealing algorithm for the traveling salesman problem
Full text link- …
