"Application of Causal Models for the Selection and Redesign of Heuristic Algorithms for Solving the Bin-Packing Problem."

by Joaquin Perez, Laura Cruz, Rodolfo Pazos, Vanesa Landero, Veronica Perez, Gerardo Reyes, Jorge Ruiz Vanoye
==================================================================================================
Reviewer: Akira IMADA, akira@bstu.by

1) How familiar are you with the subject matter of the paper? [0-5]: 4

2) Relevance to the conference [1-5]: 5

3) Contribution and originality [1-5]: 2

4) Paper readabilty, presentation and organization [1-5]: 2

5) Overall rating: 2

6) Comments to the Authors:
Authors propose a causal approach for the purpose of how we select the best algorithm among others to solve the problem given, as authors put in Instruction "Unfortunately, in real life situations, there is usually no algorithm that outperforms all the other algorithms for all instances, and therefore, the problem of selecting the best algorithm arises. Several related works have experimentally analyzed algorithm behavior in order to find the best algorithm..."

This is surely a very important target.

Then the proposed approach is tested under a set of 4 algorithms -- two variants of "threshold accepting algorithm" and two variants of "tabu-search" to solve bin packing problem. A causal model is generated for each of these four algorithms showing how good each of those is.

Here, a qustion arises. How on earth we know the best algorithm to solve the bin packing problem by applying only these 6 algorithms? 

So logic of the paper is poor, in additon to poor English expression.

This issue is strongly related to what is called No-free-lunch theory.

Also the difference from Author's previous paper [7] should be more explicitly mentioned.

7) Confidential Comments to the Editors: None