On decomposition methods in interactive user-preference based optimization

Evolutionary multi-objective optimization (EMO) methodologies have been widely applied to find a well-distributed trade-off solutions approximating to the Pareto-optimal front in the past decades. However, integrating the user-preference into the optimization to find the region of interest (ROI) [1] or preferred Pareto-optimal solutions could be more efficient and effective for the decision maker (DM) straightforwardly. In this paper, we propose several methods by combining preference-based strategy (like the reference points) with the decomposition-based multi-objective evolutionary algorithm (MOEA/D) [2], and demonstrate how preferred sets or ROIs near the different reference points specified by the DM can be found simultaneously and interactively. The study is based on the experiments conducted on a set of test problems with objectives ranging from two to fifteen objectives. Experiments have proved that the proposed approaches are more efficient and effective especially on many-objective problems to provide a set of solutions to the DM's preference, so that a better and a more reliable decision can be made.     

Preference-based evolutionary multi-objective optimization in ship weather routing

In evolutionary multi-objective optimization (EMO) the aim is to find a set of Pareto-optimal solutions. Such approach may be applied to multiple real-life problems, including weather routing (WR) of ships. The route should be optimal in terms of passage time, fuel consumption and safety of crew and cargo while taking into account dynamically changing weather conditions. Additionally it must not violate any navigational constraints (neither static nor dynamic). Since the resulting non-dominated solutions might be numerous, some user support must be provided to enable the decision maker (DM) selecting a single “best” solution. Commonly, multi-criteria decision making methods (MCDM) are utilized to achieve this goal with DM’s preferences defined a posteriori. Another approach is to apply DM’s preferences into the very process of finding Pareto-optimal solutions, which is referred to as preference-based EMO. Here the Pareto-set is limited to those solutions, which are compliant with the pre-configured user preferences. The paper presents a new tradeoff-based EMO approach utilizing configurable weight intervals assigned to all objectives. The proposed method has been applied to ship WR problem and compared with a popular reference point method: r-dominance. Presented results prove applicability and competitiveness of the proposed method to solving multi-objective WR problem.     

An interactive evolutionary multi-objective optimization and decision making procedure

With the advent of efficient techniques for multi-objective evolutionary optimization (EMO), real-world search and optimization problems are being increasingly solved for multiple conflicting objectives. During the past decade of research and application, most emphasis has been spent on finding the complete Pareto-optimal set, although EMO researchers were always aware of the importance of procedures which would help choose one particular solution from the Pareto-optimal set for implementation. This is also one of the main issues on which the classical and EMO philosophies are divided on. In this paper, we address this long-standing issue and suggest an interactive EMO procedure which will involve a decision-maker in the evolutionary optimization process and help choose a single solution at the end. This study uses many year’s of research on EMO and would hopefully encourage both practitioners and researchers to pay more attention in viewing the multi-objective optimization as a aggregate task of optimization and decision-making.     

A Case-Based Micro Interactive Genetic Algorithm (CBMIGA) for interactive learning and search: Methodology and application to groundwater monitoring design

Interactive optimization algorithms use real–time interaction to include decision maker preferences based on the subjective quality of evolving solutions. In water resources management problems where numerous qualitative criteria exist, use of such interactive optimization methods can facilitate in the search for comprehensive and meaningful solutions for the decision maker. The decision makers using such a system are, however, likely to go through their own learning process as they view new solutions and gain knowledge about the design space. This leads to temporal changes (nonstationarity) in their preferences that can impair the performance of interactive optimization algorithms. This paper proposes a new interactive optimization algorithm – Case-Based Micro Interactive Genetic Algorithm – that uses a case-based memory and case-based reasoning to manage the effects of nonstationarity in decision maker’s preferences within the search process without impairing the performance of the search algorithm. This paper focuses on exploring the advantages of such an approach within the domain of groundwater monitoring design, though it is applicable to many other problems. The methodology is tested under non-stationary preference conditions using simulated and real human decision makers, and it is also compared with a non-interactive genetic algorithm and a previous version of the interactive genetic algorithm.     

An interactive computer system for multicriteria facility location

A computer system developed to assist a decision maker in finding his most preferred efficient solution to a multicriteria location model is described. The system requires interaction between the decision maker and optimization software to conduct a heuristic search of the set of efficient solutions to the location model. A command language developed to give the user control over the system and an optimization algorithm developed for finding efficient solutions are presented.     

PROJECT Method for Multiobjective Optimization Based on Gradient Projection and Reference Points

In this paper, we propose a new interactive method for multiobjective programming (MOP) called the PROJECT method. Interactive methods in MOP are techniques that can help the decision maker (DM) to generate the most preferred solution from a set of efficient solutions. An interactive method should be capable of capturing the preferences of the DM in a pragmatic and comprehensive way. In certain decision situations, it may be easier and more reliable for DMs to follow an interactive process for providing local tradeoffs than other kinds of preferential information like aspiration levels, objective function classification, etc. The proposed PROJECT method belongs to the class of interactive local tradeoff methods. It is based on the projection of utility function gradients onto the tangent hyperplane of an efficient set and on a new local search procedure that inherits the advantages of the reference-point method to search for the best compromise solution within a local region. Most of the interactive methods based on local tradeoffs assume convexity conditions in a MOP problem, which is too restrictive in many real-life applications. The use of a reference-point procedure makes it possible to generate any efficient solutions, even the nonsupported solutions or efficient solutions located in the nonconvex part of the efficient frontier of a nonconvex MOP problem. The convergence of the proposed method is investigated. A nonlinear example is examined using the new method, as well as a case study on efficiency analysis with value judgements. The proposed PROJECT method is coded in Microsoft Visual C++ and incorporated into the software PROMOIN (Interactive MOP).     

Preference representation using Gaussian functions on a hyperplane in evolutionary multi-objective optimization

Many-objective optimization has attracted much attention in evolutionary multi-objective optimization (EMO). This is because EMO algorithms developed so far often degrade their search ability for optimization problems with four or more objectives, which are frequently referred to as many-objective problems. One of promising approaches to handle many objectives is to incorporate the preference of a decision maker (DM) into EMO algorithms. With the preference, EMO algorithms can focus the search on regions preferred by the DM, resulting in solutions close to the Pareto front around the preferred regions. Although a number of preference-based EMO algorithms have been proposed, it is not trivial for the DM to reflect his/her actual preference in the search. We previously proposed to represent the preference of the DM using Gaussian functions on a hyperplane. The DM specifies the center and spread vectors of the Gaussian functions so as to represent his/her preference. The preference handling is integrated into the framework of NSGA-II. This paper extends our previous work so that obtained solutions follow the distribution of Gaussian functions specified. The performance of our proposed method is demonstrated mainly for benchmark problems and real-world applications with a few objectives in this paper. We also show the applicability of our method to many-objective problems.     

一种引入偏好信息的多目标优化遗传算法

提出一种以交互方式引入决策者偏好信息的多目标优化遗传算法．该算法使用一种基于偏好信息的排序方法比较个体之间的优劣，并通过图形用户界面实现决策者与算法的交互．对算法的计算复杂度进行了理论分析，并进行了仿真实验．结果表明，所提算法具有较高的搜索效率，能够有效地求得期望区域内的折中解；特别当决策者的偏好发生变化时，算法能够快速做出响应，改变搜索范围，提供相应区域内的解．     

A preference multi-objective optimization based on adaptive rank clone and differential evolution

Evolutionary multi-objective optimization (EMO) algorithms have been used in various real-world applications. However, most of the Pareto domination based multi-objective optimization evolutionary algorithms are not suitable for many-objective optimization. Recently, EMO algorithm incorporated decision maker’s preferences became a new trend for solving many-objective problems and showed a good performance. In this paper, we first use a new selection scheme and an adaptive rank based clone scheme to exploit the dynamic information of the online antibody population. Moreover, a special differential evolution (DE) scheme is combined with directional information by selecting parents for the DE calculation according to the ranks of individuals within a population. So the dominated solutions can learn the information of the non-dominated ones by using directional information. The proposed method has been extensively compared with two-archive algorithm, light beam search non-dominated sorting genetic algorithm II and preference rank immune memory clone selection algorithm over several benchmark multi-objective optimization problems with from two to ten objectives. The experimental results indicate that the proposed algorithm achieves competitive results.     

Extremist vs. centrist decision behavior: quasi-convex utility functions for interactive multi-objective linear programming problems

This paper presents the fundamental theory and algorithms for identifying the most preferred alternative for a decision maker (DM) having a non-centrist (or extremist) preferential behavior. The DM is requested to respond to a set of questions in the form of paired comparison of alternatives. The approach is different than other methods that consider the centrist preferential behavior.In this paper, an interactive approach is presented to solve the multiple objective linear programming (MOLP) problem. The DM's underlying preferential function is represented by a quasi-convex value (utility) function, which is to be maximized. The method presented in this paper solves MOLP problems with quasi-convex value (utility) functions by using paired comparison of alternatives in the objective space. From the mathematical point of view, maximizing a quasi-convex (or a convex) function over a convex set is considered a difficult problem to solve, while solutions for quasi-concave (or concave) functions are currently available. We prove that our proposed approach converges to the most preferred alternative.We demonstrate that the most preferred alternative is an extreme point of the MOLP problem, and we develop an interactive method that guarantees obtaining the global most preferred alternative for the MOLP problem. This method requires only a finite number of pivoting operations using a simplex-based method, and it asks only a limited number of paired comparison questions of alternatives in the objective space. We develop a branch and bound algorithm that extends a tree of solutions at each iteration until the MOLP problem is solved. At each iteration, the decision maker has to identify the most preferred alternatives from a given subset of efficient alternatives that are adjacent extreme points to the current basis. Through the branch and bound algorithm, without asking many questions from the decision maker, all branches of the tree are implicitly enumerated until the most preferred alternative is obtained. An example is provided to show the details of the algorithm. Some computational experiments are also presented.Scope and purposeThis paper presents the fundamental theory, algorithm, and examples for identifying the most preferred alternative (solution) for a decision maker (DM) having a non-centrist (or extremist) preferential behavior for Multiple Objective Linear Programming (MOLP) problems. The DM is requested to respond to a set of questions in the form of paired comparison of alternatives.Although widely applied, Linear Programming is limited to a single objective function. In many real world situations, DMs are faced with multiple objective problems in that several competing and conflicting objectives have to be considered. For these problems, there exist many alternatives that are feasible and acceptable. However, the DM is interested in finding “the most preferred alternative”. In the past three decades, many methods have been developed for solving MOLP problems.One class of these methods is called “interactive”, in which the DM responds to a set of questions interactively so that his/her most preferred alternative can be obtained. In most of these methods, the value (utility) function (that presents the DM's preference) is assumed to be linear or additive, concave, pseudo-concave, or quasi-concave. However, for MOLP problems, there has not been any effort to recognize and solve the quasi-convex utility functions, which are among the most difficult class of problems to solve. The quasi-convex class of utility functions represents an extremist preferential behavior, while the other aforementioned methods (such as quasi-concave) represent a conservative behavioral preference. It is shown that the method converges to the optimal (the most preferred) alternative. The approach is computationally feasible for moderately sized problems.     

Optimization of testability parameters

The main objective of this study was to investigate performance models of automatic diagnostic systems taking into consideration its imperfections such as incorrect isolation and false alarms. This was accomplished by developing an optimization model to assist the decision maker in determining the optimal values of testability parameters which maximize his/her utility function. This will provide the decision maker with a tool to evaluate the parameters set forth by the designer and to assess the real capability of the diagnostic system. This tool also will help the decision maker to check if the correct detection and isolation capability of the system, as well as the imperfections of the system, e.g., incorrect isolation and false alarms are acceptable and satisfactory.

An interactive program was developed and help to implement the optimization modedl. This program provides the decision maker with enough flexibility to avaluate different different strategies and to repeat the decision process after changing one or more of the model's parameter. An example will be presented to show the application of this optimization model.     

Issues related to the computer realization of a multidisciplinary and multiobjective optimization system

Issues and novel ideas to be considered when developing computer realizations of complex multidisciplinary and multiobjective optimization systems are introduced. The aim is to discuss computer realizations that make possible both computationally efficient multidisciplinary analysis and multiobjective optimization of real world problems. We introduce software tools that make typically very time-consuming simulation processes more effective and, thus, enable even interactive multiobjective optimization with a real decision maker. In this paper, we first define a multidisciplinary and multiobjective optimization system and after that present an implementation overview of such problems including basic components participating in the solution process. Furthermore, interfaces and data flows between the components are described. A couple of important features related to the implementation are discussed in detail, for example, the usage of automatic differentiation. Finally, the ideas presented are illustrated with an industrial multiobjective optimization problem, when we describe numerical experiments related to quality properties in paper making.     

Automatic generation of controllers for embodied legged organisms: a Pareto evolutionary multi-objective approach

In this paper, we investigate the use of a self-adaptive Pareto evolutionary multi-objective optimization (EMO) approach for evolving the controllers of virtual embodied organisms. The objective of this paper is to demonstrate the trade-off between quality of solutions and computational cost. We show empirically that evolving controllers using the proposed algorithm incurs significantly less computational cost when compared to a self-adaptive weighted sum EMO algorithm, a self-adaptive single-objective evolutionary algorithm (EA) and a hand-tuned Pareto EMO algorithm. The main contribution of the self-adaptive Pareto EMO approach is its ability to produce sufficiently good controllers with different locomotion capabilities in a single run, thereby reducing the evolutionary computational cost and allowing the designer to explore the space of good solutions simultaneously. Our results also show that self-adaptation was found to be highly beneficial in reducing redundancy when compared against the other algorithms. Moreover, it was also shown that genetic diversity was being maintained naturally by virtue of the system's inherent multi-objectivity.     

用多目标演化优化算法解决约束选址问题

约束选址问题是一个多目标约束优化问题，传统算法（加权法）一次只能得到一个候选解，用多目标演化优化算法对其进行求解，可以一次得到多个候选解，给决策者提供更多的选择余地，以期获得更大的利益，数字试验表明，该方法优于传统多目标优化方法。     

Tuning of PID controller based on a multiobjective genetic algorithm applied to a robotic manipulator

Most controllers optimization and design problems are multiobjective in nature, since they normally have several (possibly conflicting) objectives that must be satisfied at the same time. Instead of aiming at finding a single solution, the multiobjective optimization methods try to produce a set of good trade-off solutions from which the decision maker may select one. Several methods have been devised for solving multiobjective optimization problems in control systems field. Traditionally, classical optimization algorithms based on nonlinear programming or optimal control theories are applied to obtain the solution of such problems. The presence of multiple objectives in a problem usually gives rise to a set of optimal solutions, largely known as Pareto-optimal solutions. Recently, Multiobjective Evolutionary Algorithms (MOEAs) have been applied to control systems problems. Compared with mathematical programming, MOEAs are very suitable to solve multiobjective optimization problems, because they deal simultaneously with a set of solutions and find a number of Pareto optimal solutions in a single run of algorithm. Starting from a set of initial solutions, MOEAs use iteratively improving optimization techniques to find the optimal solutions. In every iterative progress, MOEAs favor population-based Pareto dominance as a measure of fitness. In the MOEAs context, the Non-dominated Sorting Genetic Algorithm (NSGA-II) has been successfully applied to solving many multiobjective problems. This paper presents the design and the tuning of two PID (Proportional–Integral–Derivative) controllers through the NSGA-II approach. Simulation numerical results of multivariable PID control and convergence of the NSGA-II is presented and discussed with application in a robotic manipulator of two-degree-of-freedom. The proposed optimization method based on NSGA-II offers an effective way to implement simple but robust solutions providing a good reference tracking performance in closed loop.     

An intuitionistic fuzzy goal programming approach for finding pareto-optimal solutions to multi-objective programming problems

Multi-objective optimization in the intuitionistic fuzzy environment is the process of finding a Pareto-optimal solution that simultaneously maximizes the degree of satisfaction and minimizes the degree of dissatisfaction of an intuitionistic fuzzy decision. In this paper, a new method for solving multi-objective programming problems is developed that unlike other methods in the literature, provides compromise solutions satisfying both the conditions of intuitionistic fuzzy efficiency and Pareto-optimality. This method combines the advantages of the intuitionistic fuzzy sets concept, goal programming, and interactive procedures, and supports the decision maker in the process of solving programming problems with crisp, fuzzy, or intuitionistic fuzzy objectives and constraints. A characteristic of the proposed method is that it provides a well-structured approach for determining satisfaction and the dissatisfaction degrees that efficiently uses the concepts of violation for both objective functions and constraints. Another feature of the proposed method comes from its continuous interaction with the decision maker. In this situation, through adjusting the problem's parameters, the decision maker would have the ability of revisiting the membership and non-membership functions. Therefore, despite the lack of information at the beginning of the solving process, a compromise solution that satisfies the decision maker's preferences can be obtained. A further feature of the proposed method is the introduction of a new two-step goal programming approach for determining the compromise solutions to multi-objective problems. This approach ensures that the compromise solution obtained during each iterative step satisfies both the conditions of intuitionistic fuzzy efficiency and Pareto-optimality. The application of the proposed model is also discussed in this paper.     

基于ε占优的正交多目标差分演化算法研究

演化多目标优化是目前演化计算中热门研究方向之一.但是,要设计一种高效、鲁棒的演化多目标优化算法,使其找到接近最优和完整的非劣解集是一项很困难的任务.为了能有效求解多目标优化问题,提出了一种新的多目标差分演化算法.新算法具有如下特征:1)利用正交实验设计和连续空间量化的方法产生初始群体,使得初始群体中的个体可以均匀分布于搜索空间,并且可以使好的个体在演化过程中得到利用;2)采用Archive群体保存非劣解,并利用ε占优方法更新Archive群体,从而可以使算法较快获得分布很好的Pareto解集;3)为了加快算法收敛,提出一种基于随机选择和精英选择的混合选择机制.通过8个标准测试函数对新算法进行测试,并与其他一些多目标演化算法进行比较,其结果表明新算法可以有效逼近真实Pareto前沿且分布均匀,并且在收敛性和多样性的求解精度和稳     

An interactive rectangle elimination method for biobjective decision making

This paper presents a new man-machine interactive method for biobjective decision making. It is specifically designed to cope with both the ill-defined nature of the decision problem and the high cost of computation points in the tradeoff (Pareto optimal) set. With this method, the decision maker may efficiently approximate the tradeoff set and/or estimate his preferred objective value. First, the notion of a rectangle representation of the tradeoff set by a set of points, called experiments, and a set of rectangles, defined by the experiments, is introduced. Next, a special class of decision makers is considered. For a decision maker in this special class, the rectangle representation of the tradeoff set defines a rectangle of uncertainty that contains the decision maker's preferred objective value. A measure of the worst ease uncertainity is formulated and minimized to yield an optimal strategy for interactively selecting experiments. Finally, this strategy is employed in a general interactive algorithm that works under minimal assumptions on the tradeoff set and on the decision maker.     

An algorithm for projecting a reference direction onto thenondominated set of given points

We consider the problem of searching nondominated alternatives in a discrete multiple criteria problem. The search procedure is based on the use of a reference direction. A reference direction reflects the desire of the decision maker (DM) to specify a search direction. To find a set of given alternatives related somehow to the reference direction specified by the DRI, the reference direction has to be projected onto the set of nondominated alternatives. Our purpose is to develop an efficient algorithm for making this projection. The projection of each given reference direction determines a nondominated ordered subset. The set is provided to a decision maker for evaluation. The decision maker will choose the most preferred alternative from this subset and continues the search from this alternative with a new reference direction. The search will end when no direction of improvement is found. A critical point in the procedure is the efficiency of the projection operation. This efficiency of our algorithm is considered theoretically and numerically. The projection is made by parametrizing an achievement scalarizing function originally proposed by Wierzbicki (1980) to project any single point onto the nondominated set     

Using interactive multiobjective methods to solve DEA problems with value judgements

Data envelopment analysis (DEA) is a performance measurement tool that was initially developed without consideration of the decision maker (DM)'s preference structures. Ever since, there has been a wide literature incorporating DEA with value judgements such as the goal and target setting models. However, most of these models require prior judgements on target or weight setting. This paper will establish an equivalence model between DEA and multiple objective linear programming (MOLP) and show how a DEA problem can be solved interactively without any prior judgements by transforming it into an MOLP formulation. Various interactive multiobjective models would be used to solve DEA problems with the aid of PROMOIN, an interactive multiobjective programming software tool. The DM can then search along the efficient frontier to locate the most preferred solution where resource allocation and target levels based on the DM's value judgements can be set. An application on the efficiency analysis of retail banks in the UK is examined. Comparisons of the results among the interactive MOLP methods are investigated and recommendations on which method may best fit the data set and the DM's preferences will be made.