optimization problems involving mixed variables (i.e., variables of numerical and categorical nature) can be challenging to solve, especially in the presence of mixed-variable constraints. Moreover, when the objective...
详细信息
optimization problems involving mixed variables (i.e., variables of numerical and categorical nature) can be challenging to solve, especially in the presence of mixed-variable constraints. Moreover, when the objective function is the result of a complicated simulation or experiment, it may be expensive-to-evaluate. This paper proposes a novel surrogate-based global optimization algorithm to solve linearly constrained mixed-variable problems up to medium size (around 100 variables after encoding). The proposed approach is based on constructing a piecewise affine surrogate of the objective function over feasible samples. We assume the objective function is black-box and expensive-to-evaluate, while the linear constraints are quantifiable, unrelaxable, a priori known, and are cheap to evaluate. We introduce two types of exploration functions to efficiently search the feasible domain via mixed-integer linear programming solvers. We also provide a preference-based version of the algorithm designed for situations where only pairwise comparisons between samples can be acquired, while the underlying objective function to minimize remains unquantified. The two algorithms are evaluated on several unconstrained and constrained mixed-variable benchmark problems. The results show that, within a small number of required experiments/simulations, the proposed algorithms can often achieve better or comparable results than other existing methods.
preference-based optimization is a powerful tool to improve the performance of a system in an intuitive way. Such a methodology allows for solving optimization problems in which the decision-maker cannot evaluate the ...
详细信息
ISBN:
(纸本)9781713867890
preference-based optimization is a powerful tool to improve the performance of a system in an intuitive way. Such a methodology allows for solving optimization problems in which the decision-maker cannot evaluate the objective function related to the target problem, but rather can only express a preference such as "this is better than that" between two candidate decision vectors. In such a way, the target cost function can be easily defined and implemented, avoiding complex formulations and the use of external sources of data. Such a methodology is nowadays deeply investigated to reduce the complexity related to systems/applications tuning in real applications, providing the user with a natural procedure capable to exploit his/her knowledge and preferences. Physical human-robot collaboration (HRC) is an important field of application considering the preference-based optimization topic. Physical HRC is increasingly important in many different domains. Properly tuning the robot behavior is important to achieve satisfactory performance from the human's perspective. However, a common tuning for all the possible subjects is not feasible due to their different skills, different backgrounds, and different expected performance. However, ad hoc tuning of the robot behavior is not trivial. To address this issue, the here presented contribution applies preference-based optimization to the tuning of a physical HRC controller. In such a way, an ad hoc tuning of the robot behavior based on the perceived interaction is achieved for each subject. Experimental results show the capabilities of the method, being able to optimize the robot behavior in limited optimization trials. Copyright (c) 2022 The Authors. This is an open access article under the CC BY-NC-ND license (https://***/licenses/by-nc-nd/4.0/)
preference-based optimization algorithms are iterative procedures that seek the optimal calibration of a decision vector based only on comparisons between couples of different tunings. At each iteration, a human decis...
详细信息
preference-based optimization algorithms are iterative procedures that seek the optimal calibration of a decision vector based only on comparisons between couples of different tunings. At each iteration, a human decision-maker expresses a preference between two calibrations (samples), highlighting which one, if any, is better than the other. The optimization procedure must use the observed preferences to find the tuning of the decision vector that is most preferred by the decision-maker, while also minimizing the number of comparisons. In this work, we formulate the preference-based optimization problem from a utility theory perspective. Then, we propose GLISp-r, an extension of a recent preference-based optimization procedure called GLISp. The latter uses a Radial Basis Function surrogate to describe the tastes of the decision-maker. Iteratively, GLISp proposes new samples to compare with the best calibration available by trading off exploitation of the surrogate model and exploration of the decision space. In GLISp-r, we propose a different criterion to use when looking for new candidate samples that is inspired by MSRS, a popular procedure in the black-box optimization framework. Compared to GLISp, GLISp-r is less likely to get stuck on local optima of the preference-based optimization problem. We motivate this claim theoretically, with a proof of global convergence, and empirically, by comparing the performances of GLISp and GLISp-r on several benchmark optimization problems.
Nowadays, complex inspection processes rely heavily on human operators, while automatic systems handle simpler tasks. However, these tasks are highly repetitive and demand consistent high-quality performance throughou...
详细信息
Nowadays, complex inspection processes rely heavily on human operators, while automatic systems handle simpler tasks. However, these tasks are highly repetitive and demand consistent high-quality performance throughout, leading to significant stress for human workers. In contrast, automatic systems can help alleviate this burden. Nevertheless, configuring automatic inspection systems is challenging due to numerous parameters that require extensive time and trial-and-error adjustments. To address these issues, this project aims to introduce an optimization approach based on user preferences for configuring visual inspection systems. preference-based optimization is a potent method for enhancing system performance in an intuitive manner. This methodology enables the resolution of optimization problems when the decision-maker cannot directly assess the objective function tied to the problem at hand. Instead, they can only express preferences, such as "this option is better than that" when comparing different decision choices.
preference-based optimization is a powerful tool to improve the performance of a system in an intuitive way. Such a methodology allows for solving optimization problems in which the decision-maker cannot evaluate the ...
详细信息
preference-based optimization is a powerful tool to improve the performance of a system in an intuitive way. Such a methodology allows for solving optimization problems in which the decision-maker cannot evaluate the objective function related to the target problem, but rather can only express a preference such as “this is better than that” between two candidate decision vectors. In such a way, the target cost function can be easily defined and implemented, avoiding complex formulations and the use of external sources of data. Such a methodology is nowadays deeply investigated to reduce the complexity related to systems/applications tuning in real applications, providing the user with a natural procedure capable to exploit his/her knowledge and preferences. Physical human-robot collaboration (HRC) is an important field of application considering the preference-based optimization topic. Physical HRC is increasingly important in many different domains. Properly tuning the robot behavior is important to achieve satisfactory performance from the human's perspective. However, a common tuning for all the possible subjects is not feasible due to their different skills, different backgrounds, and different expected performance. However, ad hoc tuning of the robot behavior is not trivial. To address this issue, the here presented contribution applies preference-based optimization to the tuning of a physical HRC controller. In such a way, an ad hoc tuning of the robot behavior based on the perceived interaction is achieved for each subject. Experimental results show the capabilities of the method, being able to optimize the robot behavior in limited optimization trials.
Evolutionary multi-objective optimization (EMO) algorithms are widely used to solve problems involving multiple conflicting objectives. In general, these problems result in a well-distributed and diverse set of Pareto...
详细信息
ISBN:
(纸本)9798400704949
Evolutionary multi-objective optimization (EMO) algorithms are widely used to solve problems involving multiple conflicting objectives. In general, these problems result in a well-distributed and diverse set of Pareto-optimal solutions, consisting of individual objective-optimal solutions at their extreme and various compromise objective solutions at their core. However, in practice, decision-makers (DMs) usually have certain pre-conceived preference information which may make a majority of the Pareto solution set uninteresting to the DMs. In such cases, DM's preference information can be utilized to update EMO algorithms to focus on the preferred part of the Pareto set, rather than the entire Pareto set. While EMO researchers have proposed preference-based EMO algorithms for this purpose, appropriate metrics to evaluate their performance have received lukewarm attention. In this paper, we critically analyze a recently proposed preference-based hypervolume (R-HV) metric for its sensitivity to handle various scenarios and propose an updated version to remedy the difficulties associated with it. The updated R-HV metric is then compared with the original R-HV metric on solutions obtained from a number of preference-based EMO algorithms. The suggestion of a more appropriate R-HV metric presented in this paper should encourage further research in preference-based multi-objective optimization.
This paper proposes a method for solving optimization problems in which the decision-maker cannot evaluate the objective function, but rather can only express a preference such as "this is better than that" ...
详细信息
This paper proposes a method for solving optimization problems in which the decision-maker cannot evaluate the objective function, but rather can only express a preference such as "this is better than that" between two candidate decision vectors. The algorithm described in this paper aims at reaching the global optimizer by iteratively proposing the decision maker a new comparison to make, based on actively learning a surrogate of the latent (unknown and perhaps unquantifiable) objective function from past sampled decision vectors and pairwise preferences. A radial-basis function surrogate is fit via linear or quadratic programming, satisfying if possible the preferences expressed by the decision maker on existing samples. The surrogate is used to propose a new sample of the decision vector for comparison with the current best candidate based on two possible criteria: minimize a combination of the surrogate and an inverse weighting distance function to balance between exploitation of the surrogate and exploration of the decision space, or maximize a function related to the probability that the new candidate will be preferred. Compared to active preference learning based on Bayesian optimization, we show that our approach is competitive in that, within the same number of comparisons, it usually approaches the global optimum more closely and is computationally lighter. Applications of the proposed algorithm to solve a set of benchmark global optimization problems, for multi-objective optimization, and for optimal tuning of a cost-sensitive neural network classifier for object recognition from images are described in the paper. MATLAB and a Python implementations of the algorithms described in the paper are available at .
preference-based optimizing methods have shown their advantages and potential in exploring individual, comfortable, and effective control strategies and assistance parameters of exoskeletons during locomotion. Researc...
详细信息
ISBN:
(纸本)9798350384581;9798350384574
preference-based optimizing methods have shown their advantages and potential in exploring individual, comfortable, and effective control strategies and assistance parameters of exoskeletons during locomotion. Research indicates that compared with naive wearers, knowledgeable wearers with abundant exoskeleton assistance experience have obvious advantages in speeding up the parameters exploration process and improving the assistant performance. However, there is no existing method that could utilize the human-exoskeleton locomotion interaction experience (HELIE) to assist naive wearers during the exploration process. In this work, we propose a novel preference-based human-exoskeleton locomotion interaction experience transfer (LIET) framework, which could speed up the exploration of human-preferred parameters and acquire more satisfying results for naive wearers via the HELIE acquired from knowledgeable wearers. In addition, based on the proposed LIET framework, we establish the mathematical expression of the HELIE transfer during exoskeleton assistance. This will promote the research that concerns utilizing HELIE for exoskeleton control parameters optimizations in the future. Finally, experiments demonstrate the proposed LIET framework could speed up the exploration process and acquire more satisfying optimized results for naive wearers.
It is desirable in many multi-objective machine learning applications, such as multi-task learning with conflicting objectives, multi-objective reinforcement learning, to find a Pareto solution that can match a given ...
详细信息
It is desirable in many multi-objective machine learning applications, such as multi-task learning with conflicting objectives, multi-objective reinforcement learning, to find a Pareto solution that can match a given preference of a decision maker. These problems are often large-scale with available gradient information but cannot be handled very well by the existing algorithms. To tackle this issue, this paper proposes a novel predict-and-correct framework for locating a Pareto solution that fits the preference of a decision maker. In the proposed framework, a constraint function is introduced in the search progress to align the solution with a user-specific preference, which can be optimized simultaneously with multiple objective functions. Experimental results show that our proposed method can efficiently find a particular Pareto solution under the demand of a decision maker for standard multiobjective benchmark, multi-task learning, multi-objective reinforcement learning problems with more than thousands of decision variables.
One of the main objectives of the fifth industrial revolution is the design and implementation of human-centric production environments. The human is, indeed, placed in the center of the production environment, having...
详细信息
One of the main objectives of the fifth industrial revolution is the design and implementation of human-centric production environments. The human is, indeed, placed in the center of the production environment, having a supervision/leading role instead of carrying out heavy/repetitive tasks. To enhance such an industrial paradigm change, industrial operators have to be provided with the tools they need to naturally and easily transfer their knowledge to robotic systems. Such expertise, in fact, is difficult to be coded, especially for non-expert programmers. In addition, due to the reduced specialized manpower, the capability to transfer such knowledge into robotic systems is becoming increasingly critical and demanding. In response to this need, this contribution aims to propose and validate a human-centric approach to transfer the human's knowledge of a task into the robot controller making use of qualitative feedback only (to this end, preference-based optimization is employed). In addition, the modeled human's knowledge is exploited by an optimization algorithm (i.e., nonlinear programming) to maximize the task performance while managing the task constraints. The proposed approach has been implemented and validated for a robotic sealant material deposition task. On the basis of the qualitative feedback provided by the operator, the knowledge related to the deposition task has been transferred to the robotic system and optimized to deal with the hardware and task constraints. The achieved results show the generalization of the approach, making it possible to optimize the deposition task output.
暂无评论