Partial evaluation in Rank Aggregation Problems

Partial Evaluation and Efficient Discarding for the Maximal Covering Location Problem

“…Constraint (2) implies that a demand node i is covered only when there is one or more open facilities that belongs to the set N i . Constraint (3) implies that the number of facilities to be located is p. Fig. 1 shows a solution for a MCLP instance with p = 4, which is a binary vector where each one (1) indicates an open facility, and 0 otherwise.…”

“…One way to reduce this computational cost is by using surrogate functions [2], which may be either approximate or exact versions of the original objective function of the problem. Two approaches to obtain equivalent (not approximate) versions of objective function are: partial evaluation [3] and efficient discarding [4]. Partial evaluation takes into account the influence of the operators on the quality of the solutions.…”

“…Partial evaluation takes into account the influence of the operators on the quality of the solutions. It only evaluates the part of the objective function that was modified by the action of the operators [3]. On the other hand, efficient discarding consists in stopping the evaluation when it is known that the remaining part of the solution will not affect the final value of the objective function.…”

“…Both approaches have been successfully used in several problems, The associate editor coordinating the review of this manuscript and approving it for publication was Xujie Li . such as: aggregation of rankings [3], graph drawing [5], task planning [4] and binary problems [6]. These examples indicate the growing interest in the objective function cost reduction in metaheuristics.…”

Median constrained bucket order rank aggregation

“…The rank aggregation problem has been approached using several different approaches, some of which work only when complete rankings are used as the input and produce an output that is a full con sensus ranking (Meila et al, 2007;Aledo et al, 2013;D'Ambrosia et al, 2015). Other proposals work with complete, tied and partial rankings and produce an out put solution that either can contain ties (Emond and Mason, 2002;Gionis et al, 2006;Lin and Ding, 2009;Ukkonen et al, 2009;Lin, 2010;Amodio et al, 2016;D'Ambrosia et al, 2017;Aledo et al, 2017b) or cannot contain ties ( Aledo et al, 2017a;Badal and Das, 2018). By following the classification made by Cook (2006), there are two broad classes of approaches to consensus ranking: the so-called ad hoc methods, which are generally based on counting such as Borda or Condorcet-like tools, and the distance-based approaches, for which the detection of the consensus ranking is based on the minimization of a distance measure that is suitably defined for preference rankings.…”