Hassene Aissi

Min–max and min–max regret versions of combinatorial optimization problems: A survey

Min-max and min-max regret criteria are commonly used to define robust solutions. After motivating the use of these criteria, we present general results. Then, we survey complexity results for the min-max and min-max regret versions of some combinatorial optimization problems: shortest path, spanning tree, assignment, min cut, min s-t cut, knapsack. Since most of these problems are NP-hard, we also investigate the approximability of these problems. Furthermore, we present algorithms to solve these problems to optimality.

Complexity of the min-max and min-max regret assignment problems

This paper investigates the complexity of the min-max and min-max regret assignment problems both in the discrete scenario and interval data cases. We show that these problems are strongly NP-hard for an unbounded number of scenarios. We also show that the interval data min-max regret assignment problem is strongly NP-hard.

Approximation of min-max and min-max regret versions of some combinatorial optimization problems

This paper investigates, for the first time in the literature, the approximation of min–max (regret) versions of classical problems like shortest path, minimum spanning tree, and knapsack. For a constant number of scenarios, we establish fully polynomial-time approximation schemes for the min–max versions of these problems, using relationships between multi-objective and min–max optimization. Using dynamic programming and classical trimming techniques, we construct a fully polynomial-time approximation scheme for min–max regret shortest path. We also establish a fully polynomial-time approximation scheme for min–max regret spanning tree and prove that min–max regret knapsack is not at all approximable. For a non-constant number of scenarios, in which case min–max and min–max regret versions of polynomial-time solvable problems usually become strongly NP-hard, non-approximability results are provided for min–max (regret) versions of shortest path and spanning tree.

Approximation complexity of min-max (regret) versions of shortest path, spanning tree, and knapsack

This paper investigates, for the first time in the literature, the approximation of min-max (regret) versions of classical problems like shortest path, minimum spanning tree, and knapsack. For a bounded number of scenarios, we establish fully polynomial-time approximation schemes for the min-max versions of these problems, using relationships between multi-objective and min-max optimization. Using dynamic programming and classical trimming techniques, we construct a fully polynomial-time approximation scheme for min-max regret shortest path. We also establish a fully polynomial-time approximation scheme for min-max regret spanning tree and prove that min-max regret knapsack is not at all approximable. We also investigate the case of an unbounded number of scenarios, for which min-max and min-max regret versions of polynomial-time solvable problems usually become strongly NP-hard. In this setting, non-approximability results are provided for min-max (regret) versions of shortest path and spanning tree.

Complexity of the min-max (regret) versions of cut problems

This paper investigates the complexity of the min-max and min-max regret versions of the s–t min cut and min cut problems. Even if the underlying problems are closely related and both polynomial, we show that the complexity of their min-max and min-max regret versions, for a constant number of scenarios, are quite contrasted since they are respectively strongly NP-hard and polynomial. Thus, we exhibit the first polynomial problem, s–t min cut, whose min-max (regret) versions are strongly NP-hard. Also, min cut is one of the few polynomial problems whose min-max (regret) versions remain polynomial. However, these versions become strongly NP-hard for a non constant number of scenarios. In the interval data case, min-max versions are trivially polynomial. Moreover, for min-max regret versions, we obtain the same contrasted result as for a constant number of scenarios: min-max regret s–t cut is strongly NP-hard whereas min-max regret cut is polynomial.

Approximating min-max (regret) versions of some polynomial problems

While the complexity of min-max and min-max regret versions of most classical combinatorial optimization problems has been thoroughly investigated, there are very few studies about their approximation. For a bounded number of scenarios, we establish a general approximation scheme which can be used for min-max and min-max regret versions of some polynomial problems. Applying this scheme to shortest path and minimum spanning tree, we obtain fully polynomial-time approximation schemes with much better running times than the ones previously presented in the literature.

Minimizing the number of late jobs on a single machine under due date uncertainty

We study the problem of minimizing the number of late jobs on a single machine where job processing times are known precisely and due dates are uncertain. The uncertainty is captured through a set of scenarios. In this environment, an appropriate criterion to select a schedule is to find one with the best worst-case performance, which minimizes the maximum number of late jobs over all scenarios. For a variable number of scenarios and two distinct due dates over all scenarios, the problem is proved NP-hard in the strong sense and non-approximable in pseudo-polynomial time with approximation ratio less than 2. It is polynomially solvable if the number s of scenarios and the number v of distinct due dates over all scenarios are given constants. An O(nlog n) time s-approximation algorithm is suggested for the general case, where n is the number of jobs, and a polynomial 3-approximation algorithm is suggested for the case of unit-time jobs and a constant number of scenarios. Furthermore, an O(ns+v−2/(v−1)v−2) time dynamic programming algorithm is presented for the case of unit-time jobs. The problem with unit-time jobs and the number of late jobs not exceeding a given constant value is solvable in polynomial time by an enumeration algorithm. The obtained results are related to a min-max assignment problem, an exact assignment problem and a multi-agent scheduling problem.

Robustness in Multi-criteria Decision Aiding

After bringing precisions to the meaning we give to several of the terms used in this chapter (e.g., robustness, result, procedure, method, etc.), we highlight the principal characteristics of most of the publications about robustness. Subsequently, we present several partial responses to the question, “Why is robustness a matter of interest in Multi-Criteria Decision Aiding (MCDA)?” (see Section 4.2). Only then do we provide an outline for this chapter. At this point, we introduce the concept of variable setting, which serves to connect what we define as the formal representation of the decision-aiding problem and the real-life decisional context. We then introduce five typical problems that will serve as reference problems in the rest of the chapter. Section 4.3 deals with recent approaches that involve a single robustness criterion completing (but not replacing) a preference system that has been defined previously, independently of the robustness concern. The following section deals with approaches in which the robustness concern is modelled using several criteria. Section 4.5 deals with the approaches in which robustness is considered other than by using one or several criteria to compare the solutions. These approaches generally involve using one or several properties destined to characterize the robust solution or to draw robust conclusions. In the last three sections, in addition to describing the appropriate literature, we suggest some avenues for new development and in some cases, we present some new approaches.

General approximation schemes for min-max (regret) versions of some (pseudo-)polynomial problems

While the complexity of min-max and min-max regret versions of most classical combinatorial optimization problems has been thoroughly investigated, there are very few studies about their approximation. For a bounded number of scenarios, we establish general approximation schemes which can be used for min-max and min-max regret versions of some polynomial or pseudo-polynomial problems. Applying these schemes to shortest path, minimum spanning tree, minimum weighted perfect matching on planar graphs, and knapsack problems, we obtain fully polynomial-time approximation schemes with better running times than the ones previously presented in the literature.

Complexity of the min-max (regret) versions of min cut problems

This paper investigates the complexity of the min-max and min-max regret versions of the min s-t cut and min cut problems. Even if the underlying problems are closely related and both polynomial, the complexities of their min-max and min-max regret versions, for a constant number of scenarios, are quite contrasted since they are respectively strongly NP-hard and polynomial. However, for a non-constant number of scenarios, these versions become strongly NP-hard for both problems. In the interval scenario case, min-max versions are trivially polynomial. Moreover, for min-max regret versions, we obtain the same contrasted results as for a constant number of scenarios: min-max regret min s-t cut is strongly NP-hard whereas min-max regret min cut is polynomial.