Goal Programming: realistic targets for the near future

Abstract

Goal Programming (GP) can be regarded as one of the most widely used multicriteria decision-making techniques. In this paper, two surveys are carried out. First, the evolution of GP since its birth to the present time, in terms of number of publications, references, journals, etc., has been studied. Second, a more in-depth survey has been carried out, which covers the publications from year 2000 to the present time. All the references are listed, and some conclusions and future research lines have been extracted about the late trends of GP.

Multiple Criteria Decision Making, Multiattribute Utility Theory: Recent Accomplishments and What Lies Ahead

“…Approaches for solving multiple criteria optimization problems fall into the following categories: interactive methods, goal programming, vector-maximum algorithms, and evolutionary procedures, although these categories are not necessarily mutually exclusive. For technical discussions of these and other specific MCDM/MAUT methods, see for example, Edwards et al(2007), Figueira et al (2005, Ignizio and Romero (2003), and Caballero et al (2007).…”

Addressing uncertainty in sugarcane harvest planning through a revised multi-choice goal programming model

“…In this context, and in order to handle a multiple objective situation, GP models have been applied with success [32] to solve optimization models under uncertainty [33].…”

A particle swarm solution based on lexicographical goal programming for a multiobjective fuzzy open shop problem

“…Pareto optimality is undoubtedly the most extended approach to multicriteria optimisation but, quoting [12], "it is not the end of the story". Among the alternative approaches, lexicographic and goal programming methods are some of the most popular ones [6]. The philosophy behind goal programming can be traced back to the theories of rational decision developed in the 1950s, especially the concept of satisficing solutions: in a complex environment, the decision maker's aim may be to reach a certain satisfactory level for every relevant objective, rather than optimising its value [34].…”

“…1. This polygon is determined on the left by the starting time (4,5,7) and on the right by the completion time (6,8,10); notice that in the case that starting and completion times were real numbers, the polygon would turn into a rectangle, which is the standard way of representing task execution times in deterministic Gantt charts. The makespan can be easily calculated as C max = (9,12,17), so E[C max ] = 12.5.…”