Strange: An interactive method for multi-objective linear programming under uncertainty

doi:10.1016/0377-2217(86)90160-8

European Journal of Operational Research

Volume 26, Issue 1, July 1986, Pages 65-82

https://doi.org/10.1016/0377-2217(86)90160-8 Get rights and content

Abstract

In the field of investment planning within a time horizon, problems typically involve multiple objectives, and basic data are uncertain. In a large number of cases, these decision problems can be written as linear programming problems in which time dependent uncertainties affect the coefficients and the right hand side of constraints. Given the possibility of defining plausible scenarios on basic data, discrete sets of such coefficients are given, each with its subjective probability of occurrence. The corresponding structure is then characteristic for Multi-Objective Stochastic Linear Programming (MOSLP).

In the paper, an interactive procedure is described to obtain a best compromise for such a MOSLP problem. This algorithm, called Strange, extends the Stem method to take the random aspects into account. It involves in particular, the concepts of stochastic programming with recourse. In its interactive steps, the efficiency projection techniques are used to provide the decision-maker with detailed graphical information on efficient solution families.

As an illustration of the successive steps, a didactic example is solved in some detail, and the results of a case study in energy planning are given.

References (13)

R. Benayoun et al.
Linear programming with multiple objective functions: Step method (Stem)
Mathematical Programming
(1971)
C. Carlsson
Solving ill-structured problems through well-structured fuzzy programming
M. Despontin
Interactive economic policy formulation with multiregional econometric models
A. Goicoecha et al.
Introduction to Multi-Objective Analysis with Engineering and Business Application
(1982)
P. Kall
Stochastic Linear Programming
(1976)
P. Kunsch et al.
Nuclear fuel cycle optimization using multi-objective stochastic linear programming

There are more references available in the full text version of this article.

Cited by (114)

New complex fuzzy multiple objective programming procedure for a portfolio making under uncertainty
2020, Applied Soft Computing Journal
Citation Excerpt :
Newer interactive methods can be seen in Korhonen and Wallenius [42]. In the 1980s, the stochastic interactive method began to be developed, e.g. by Teghem et al. [43]. The modern stochastic interactive approaches are from Balibek and Köksalan [44], Köksalan and Şakar [45] and Wang et al. [46].
This article deals with a still current decision making problem — where to invest and how much. A few “tools” can help to answer this question, namely fundamental, technical or psychological analysis. “Soft” aspects as a human intuition or mood on the capital market can also play a nonnegligible role. These approaches and techniques (mainly individually) provide a partial information on investment reality. In practice, an investment decision, or portfolio, is mostly made just based on such a limited information. Moreover, the assets’ shares in the portfolio must be implicitly added up. This “crumbled” approach may not be an expected support to make a satisfactory investment decision. Therefore, a complex decision making procedure is practically designed. It is based on the fuzzy multiple objective programming method whose unique algorithm, working with triangular fuzzy numbers, is proposed in this article. Although these methods of decision making theory are not widely used on the capital market, they can be very convenient. The developed approach enables multiple objectives to be simultaneously accepted, as well as vague information about the assets or vague investor preferences. Moreover, a valuable output of this multi-factor procedure is an explicit quantification of the assets’ share in the portfolio. The ability to include all these aspects, which plays a significant role in a decision making on the capital market, proves a strong application power (and advantage over the mentioned concepts) to make an investment portfolio satisfactorily. This strength is demonstrated on the Czech capital market with open unit trusts offered by Česká spořitelna. Based on the specified investment strategy, the satisfactory investment portfolio consisting of four open unit trusts is effectively made.
Pareto uncertainty index for evaluating and comparing solutions for stochastic multiple objective problems
2020, European Journal of Operational Research
Many engineering problems have multiple conflicting objectives, and they are also stochastic due to inherent uncertainties. One way to represent the multi-objective nature of problems is to use the Pareto optimality to show the trade-off between objectives. Pareto optimality involves the identification of solutions that are not dominated by other solutions based on their respective objective functions. However, the Pareto optimality concept does not contain any information about the uncertainty of solutions. Evaluation and comparison of solutions becomes difficult when the objective functions are subjected to uncertainty. A new metric, the Pareto Uncertainty Index (PUI), is presented. This metric includes uncertainty due to the stochastic coefficients in the objective functions as part of the Pareto optimality concept to form an extended probabilistic Pareto set, we define as the p-Pareto set. The decision maker can observe and assess the randomness of solutions and compare the promising solutions according to their performance of satisfying objectives and any undesirable uncertainty. The PUI is an effective and convenient decision-making tool to compare promising solutions with multiple uncertain objectives.
Scenario-based multi-period program optimization for capability-based planning using evolutionary algorithms
2017, Applied Soft Computing Journal
Capability-based planning (CBP) is a strategy focused planning framework that facilitates organizations to systematically develop capacity to achieve their business objectives in highly uncertain, dynamic and competitive environments. Capability programming is an integral part of CBP which requires selecting a portfolio of capability projects for execution, referred as a capability program, such that the overall strategic risk facing the planning organization across a number of projected future operating scenarios is minimized while maintaining the most economical choice. It is a challenging optimization problem that requires handling a number of dynamic constraints and objectives that vary throughout the entire planning horizon. An optimizing simulation approach is presented in this paper that combines an evolutionary multi-objective optimization algorithm with a reinforcement learning technique to generate capability programs which optimize strategic risks and program costs across multiple planning scenarios as well as over a rolling planning horizon. The role of the optimization algorithm in this approach is to search for the non-dominated capability programs at each decision point by minimizing the strategic risks associated with individual capability projects across a number of planning scenarios as well as the total cost of the program. The reinforcement learning algorithm, on the other hand, searches horizontally within the set of non-dominated programs to minimize capability risks and costs over the entire planning horizon. The methodology is evaluated on a test problem generated based on the data distributions in an Australian Defence Capability Plan and the performance is compared with two myopic heuristic methods.
Bayesian interval robust optimization for sustainable energy system planning in Qiqihar City, China
2016, Energy Economics
Old industrial bases in Northeastern China have been experiencing a series of problems such as energy scarcity, economic slack and environmental pollution due to lack of scientific planning of energy systems. Effective energy system management is desirable for avoiding occurrence of these problems in the following decades under the national policy of revitalizing these industrial bases. This task is challenged by system complexity, data unavailability, and inaccurate demand forecasting which can hardly be resolved in existing studies. For improving energy system management in Qiqihar City, a representative industrial base in Northeastern China, under these challenges, a large-scale and fine-resolution Bayesian interval robust energy system optimization (BIRESO) approach is developed in this study. In detail, the structure, characteristics, problems and complexities of the energy system in this city are identified and analyzed. Based on these efforts, a series of optimized energy system management schemes are generated through construction of a BIRESO model and development of a solution algorithm. The obtained solutions can be ensured to remain robust when input data changes through controlling the degree of solution conservatism in accordance with probabilistic bounds on constraint violation. Furthermore, Bayesian estimation method was employed for effectively achieving dynamic forecast of energy demands. The tradeoff among energy security, economic development and environmental conservation under multiple uncertainties is revealed, and the optimal balance among them is identified. A scenario analysis approach is used to quantify the effects of economic growth on energy system schemes based on several developed scenarios. Results show that the BIRESO approach is capable of providing scientific support for energy supply planning, renewable energy development, technological advancement, facility expansion, energy allocation, pollution control, economic impacts assessment, and other energy related activities in Qiqihar City.
Fuzzy decision making for multiobjective stochastic programming problems
2016, Fuzzy Sets and Systems
Citation Excerpt :
These approaches were extended to multiobjective stochastic programming problems (MOSPs) by using a goal programming technique [3]. An interactive method called STRANGE was proposed for MOSPs with discrete random variables [18]. Caballero et al. [1] and Stancu-Minasian [16] defined the Pareto optimality concept for MOSPs on the basis of a probability maximization model and a fractile optimization model.
In this study, we propose a fuzzy decision making method for multiobjective stochastic linear programming problems with variance covariance matrices (MOSLPs), in which the criteria of probability maximization and fractile optimization are simultaneously considered. For a probability maximization model for an MOSLP, the decision maker is required to specify permissible objective levels. However, fewer values of permissible objective levels for minimization problems result in fewer values of the corresponding distribution function because of the conflicts between them. Similarly, for a fractile optimization model for an MOSLP, the decision maker is required to specify permissible probability levels. However, due to conflicts between the probability levels and the corresponding objective functions, larger values of permissible probability levels result in larger values of the corresponding objective function for minimization problems. In this study, it is assumed that the decision maker has fuzzy goals not only for permissible objective levels of a probability maximization model but also for permissible probability levels of a fractile optimization model, and such fuzzy goals are quantified by eliciting the corresponding membership functions. On the basis of the fuzzy decision, the satisfactory solution of the decision maker is obtained by applying a convex programming technique.
A novel approach for multi-objective linear programming model under spherical fuzzy environment and its application
2024, Journal of Intelligent and Fuzzy Systems

View all citing articles on Scopus

View full text

Strange: An interactive method for multi-objective linear programming under uncertainty

Abstract

Linear programming with multiple objective functions: Step method (Stem)

Mathematical Programming

Solving ill-structured problems through well-structured fuzzy programming

Interactive economic policy formulation with multiregional econometric models

Introduction to Multi-Objective Analysis with Engineering and Business Application

Stochastic Linear Programming

Nuclear fuel cycle optimization using multi-objective stochastic linear programming