YZ
Yingqian Zhang
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
8 records found
1
Journal article
(2025)
-
Jesse van Remmerden, Maurice Kenter, Diederik M. Roijers, Charalampos Andriotis, Yingqian Zhang, Zaharah Bukhsh
In this paper, we introduce multi-objective deep centralized multi-agent actor-critic (MO-DCMAC), a multi-objective reinforcement learning method for infrastructural maintenance optimization, an area traditionally dominated by single-objective reinforcement learning (RL) approaches. Previous single-objective RL methods combine multiple objectives, such as probability of collapse and cost, into a singular reward signal through reward-shaping. In contrast, MO-DCMAC can optimize a policy for multiple objectives directly, even when the utility function is nonlinear. We evaluated MO-DCMAC using two utility functions, which use probability of collapse and cost as input. The first utility function is the threshold utility, in which MO-DCMAC should minimize cost so that the probability of collapse is never above the threshold. The second is based on the failure mode, effects, and criticality analysis methodology used by asset managers to assess maintenance plans. We evaluated MO-DCMAC, with both utility functions, in multiple maintenance environments, including ones based on a case study of the historical quay walls of Amsterdam. The performance of MO-DCMAC was compared against multiple rule-based policies based on heuristics currently used for constructing maintenance plans. Our results demonstrate that MO-DCMAC outperforms traditional rule-based policies across various environments and utility functions.
...
In this paper, we introduce multi-objective deep centralized multi-agent actor-critic (MO-DCMAC), a multi-objective reinforcement learning method for infrastructural maintenance optimization, an area traditionally dominated by single-objective reinforcement learning (RL) approaches. Previous single-objective RL methods combine multiple objectives, such as probability of collapse and cost, into a singular reward signal through reward-shaping. In contrast, MO-DCMAC can optimize a policy for multiple objectives directly, even when the utility function is nonlinear. We evaluated MO-DCMAC using two utility functions, which use probability of collapse and cost as input. The first utility function is the threshold utility, in which MO-DCMAC should minimize cost so that the probability of collapse is never above the threshold. The second is based on the failure mode, effects, and criticality analysis methodology used by asset managers to assess maintenance plans. We evaluated MO-DCMAC, with both utility functions, in multiple maintenance environments, including ones based on a case study of the historical quay walls of Amsterdam. The performance of MO-DCMAC was compared against multiple rule-based policies based on heuristics currently used for constructing maintenance plans. Our results demonstrate that MO-DCMAC outperforms traditional rule-based policies across various environments and utility functions.
The first AI4TSP competition
Learning to solve stochastic routing problems
Journal article
(2023)
-
Yingqian Zhang, Laurens Bliek, Paulo da Costa, Reza Refaei Afshar, Robbert Reijnen, Tom Catshoek, Daniël Vos, Sicco Verwer, Fynn Schmitt-Ulms, More Authors...
This paper reports on the first international competition on AI for the traveling salesman problem (TSP) at the International Joint Conference on Artificial Intelligence 2021 (IJCAI-21). The TSP is one of the classical combinatorial optimization problems, with many variants inspired by real-world applications. This first competition asked the participants to develop algorithms to solve an orienteering problem with stochastic weights and time windows (OPSWTW). It focused on two learning approaches: surrogate-based optimization and deep reinforcement learning. In this paper, we describe the problem, the competition setup, and the winning methods, and give an overview of the results. The winning methods described in this work have advanced the state-of-the-art in using AI for stochastic routing problems. Overall, by organizing this competition we have introduced routing problems as an interesting problem setting for AI researchers. The simulator of the problem has been made open-source and can be used by other researchers as a benchmark for new learning-based methods. The instances and code for the competition are available at https://github.com/paulorocosta/ai-for-tsp-competition.
...
This paper reports on the first international competition on AI for the traveling salesman problem (TSP) at the International Joint Conference on Artificial Intelligence 2021 (IJCAI-21). The TSP is one of the classical combinatorial optimization problems, with many variants inspired by real-world applications. This first competition asked the participants to develop algorithms to solve an orienteering problem with stochastic weights and time windows (OPSWTW). It focused on two learning approaches: surrogate-based optimization and deep reinforcement learning. In this paper, we describe the problem, the competition setup, and the winning methods, and give an overview of the results. The winning methods described in this work have advanced the state-of-the-art in using AI for stochastic routing problems. Overall, by organizing this competition we have introduced routing problems as an interesting problem setting for AI researchers. The simulator of the problem has been made open-source and can be used by other researchers as a benchmark for new learning-based methods. The instances and code for the competition are available at https://github.com/paulorocosta/ai-for-tsp-competition.
Journal article
(2021)
-
Qing Chuan Ye, Jason. S. Rhuggenaath, Yingqian Zhang, Sicco Verwer, Michiel Jurgen Hilgeman
Designing auction parameters for online industrial auctions is a complex problem due to highly heterogeneous items. Currently, online auctioneers rely heavily on their experts in auction design. The ability of predicting how well an auction will perform prior to the start comes in handy for auctioneers. If an item is expected to be a low-performing item, the auctioneer can take certain actions to influence the auction outcome. For instance, the starting selling price of the item can be modified, or the location where the item is displayed on the website can be changed to attract more attention. In this paper, we take a real-world industrial auction data set and investigate how we can improve upon the expert’s design using insights learned from data. More specifically, we first construct a classification model that predicts the expected performance of auctions. We propose a data driven auction design framework (called DDAD) that combines the expert’s knowledge with the learned prediction model, in order to find the best parameter values, i.e., starting price and display positions of the items, for a given new auction. The prediction model is evaluated, and the new design for several auctions is discussed and validated with the auction experts.
...
Designing auction parameters for online industrial auctions is a complex problem due to highly heterogeneous items. Currently, online auctioneers rely heavily on their experts in auction design. The ability of predicting how well an auction will perform prior to the start comes in handy for auctioneers. If an item is expected to be a low-performing item, the auctioneer can take certain actions to influence the auction outcome. For instance, the starting selling price of the item can be modified, or the location where the item is displayed on the website can be changed to attract more attention. In this paper, we take a real-world industrial auction data set and investigate how we can improve upon the expert’s design using insights learned from data. More specifically, we first construct a classification model that predicts the expected performance of auctions. We propose a data driven auction design framework (called DDAD) that combines the expert’s knowledge with the learned prediction model, in order to find the best parameter values, i.e., starting price and display positions of the items, for a given new auction. The prediction model is evaluated, and the new design for several auctions is discussed and validated with the auction experts.
Solving bin-packing problems under privacy preservation
Possibilities and trade-offs
Journal article
(2019)
-
Rowan Hoogervorst, Yingqian Zhang, Gamze Tillem, Zekeriya Erkin, Sicco Verwer
We investigate the trade-off between privacy and solution quality that occurs when a k-anonymized database is used as input to the bin-packing optimization problem. To investigate the impact of the chosen anonymization method on this trade-off, we consider two recoding methods for k-anonymity: full-domain generalization and partition-based single-dimensional recoding. To deal with the uncertainty created by anonymization in the bin-packing problem, we utilize stochastic programming and robust optimization methods. Our computational results show that the trade-off is strongly dependent on both the anonymization and optimization method. On the anonymization side, we see that using single dimensional recoding leads to significantly better solution quality than using full domain generalization. On the optimization side, we see that using stochastic programming, where we use the multiset of values in an equivalence class, considerably improves the solutions. While publishing these multisets makes the database more vulnerable to a table linkage attack, we argue that it is up to the data publisher to reason if such a loss of anonymization weighs up to the increase in optimization performance.
...
We investigate the trade-off between privacy and solution quality that occurs when a k-anonymized database is used as input to the bin-packing optimization problem. To investigate the impact of the chosen anonymization method on this trade-off, we consider two recoding methods for k-anonymity: full-domain generalization and partition-based single-dimensional recoding. To deal with the uncertainty created by anonymization in the bin-packing problem, we utilize stochastic programming and robust optimization methods. Our computational results show that the trade-off is strongly dependent on both the anonymization and optimization method. On the anonymization side, we see that using single dimensional recoding leads to significantly better solution quality than using full domain generalization. On the optimization side, we see that using stochastic programming, where we use the multiset of values in an equivalence class, considerably improves the solutions. While publishing these multisets makes the database more vulnerable to a table linkage attack, we argue that it is up to the data publisher to reason if such a loss of anonymization weighs up to the increase in optimization performance.
Conference paper
(2018)
-
Jason. S. Rhuggenaath, Yingqian Zhang, Alp Akcay, Uzay Kaymak, Sicco Verwer
A popular method in machine learning for super-vised classification is a decision tree. In this work we propose a new framework to learn fuzzy decision trees using mathematical programming. More specifically, we encode the problem of constructing fuzzy decision trees using a Mixed Integer Linear Programming (MIP) model, which can be solved by any optimization solver. We compare the performance of our method with the performance of off-the-shelf decision tree algorithm CART and Fuzzy Inference Systems (FIS) using benchmark data-sets. Our initial results are promising and show the advantages of using non-crisp boundaries for improving classification accuracy on testing data.
...
A popular method in machine learning for super-vised classification is a decision tree. In this work we propose a new framework to learn fuzzy decision trees using mathematical programming. More specifically, we encode the problem of constructing fuzzy decision trees using a Mixed Integer Linear Programming (MIP) model, which can be solved by any optimization solver. We compare the performance of our method with the performance of off-the-shelf decision tree algorithm CART and Fuzzy Inference Systems (FIS) using benchmark data-sets. Our initial results are promising and show the advantages of using non-crisp boundaries for improving classification accuracy on testing data.
In a sequential auction with multiple bidding agents, the problem of determining the ordering of the items to sell in order to maximize the expected revenue is highly challenging. The challenge is largely due to the fact that the autonomy and private information of the agents heavily influence the outcome of the auction.
The main contribution of this paper is two-fold. First, we demonstrate how to apply machine learning techniques to solve the optimal ordering problem in sequential auctions. We learn regression models from historical auctions, which are subsequently used to predict the expected value of orderings for new auctions. Given the learned models, we propose two types of optimization methods: a black-box best-first search approach, and a novel white-box approach that maps learned regression models to integer linear programs (ILP), which can then be solved by any ILP-solver. Although the studied auction design problem is hard, our proposed optimization methods obtain good orderings with high revenues.
Our second main contribution is the insight that the internal structure of regression models can be efficiently evaluated inside an ILP solver for optimization purposes. To this end, we provide efficient encodings of regression trees and linear regression models as ILP constraints. This new way of using learned models for optimization is promising. As the experimental results show, it significantly outperforms the black-box best-first search in nearly all settings.
...
In a sequential auction with multiple bidding agents, the problem of determining the ordering of the items to sell in order to maximize the expected revenue is highly challenging. The challenge is largely due to the fact that the autonomy and private information of the agents heavily influence the outcome of the auction.
The main contribution of this paper is two-fold. First, we demonstrate how to apply machine learning techniques to solve the optimal ordering problem in sequential auctions. We learn regression models from historical auctions, which are subsequently used to predict the expected value of orderings for new auctions. Given the learned models, we propose two types of optimization methods: a black-box best-first search approach, and a novel white-box approach that maps learned regression models to integer linear programs (ILP), which can then be solved by any ILP-solver. Although the studied auction design problem is hard, our proposed optimization methods obtain good orderings with high revenues.
Our second main contribution is the insight that the internal structure of regression models can be efficiently evaluated inside an ILP solver for optimization purposes. To this end, we provide efficient encodings of regression trees and linear regression models as ILP constraints. This new way of using learned models for optimization is promising. As the experimental results show, it significantly outperforms the black-box best-first search in nearly all settings.
We encode the problem of learning the optimal decision tree of a given depth as an integer optimization problem. We show experimentally that our method (DTIP) can be used to learn good trees up to depth 5 from data sets of size up to 1000. In addition to being efficient, our new formulation allows for a lot of flexibility. Experiments show that we can use the trees learned from any existing decision tree algorithms as starting solutions and improve the trees using DTIP. Moreover, the proposed formulation allows us to easily create decision trees with different optimization objectives instead of accuracy and error, and constraints can be added explicitly during the tree construction phase. We show how this flexibility can be used to learn discrimination-aware classification trees, to improve learning from imbalanced data, and to learn trees that minimise false positive/negative errors.
...
We encode the problem of learning the optimal decision tree of a given depth as an integer optimization problem. We show experimentally that our method (DTIP) can be used to learn good trees up to depth 5 from data sets of size up to 1000. In addition to being efficient, our new formulation allows for a lot of flexibility. Experiments show that we can use the trees learned from any existing decision tree algorithms as starting solutions and improve the trees using DTIP. Moreover, the proposed formulation allows us to easily create decision trees with different optimization objectives instead of accuracy and error, and constraints can be added explicitly during the tree construction phase. We show how this flexibility can be used to learn discrimination-aware classification trees, to improve learning from imbalanced data, and to learn trees that minimise false positive/negative errors.