**Tool Switching Problem (ToSP) **

ToSP
involves scheduling a number of jobs on a single machine such that the
resulting number of tool switches required is kept to a minimum.

This can be
formalized as follows: let a ToSP instance be
represented by a 4-tuple, ** I=(C, n, m, A)** where,

·
** C**
denotes the magazine capacity (i.e., number of available slots),

·
** n** is the
number of jobs to be processed,

·
** m** is the total number of tools required to
process all jobs (it is assumed that

·
** A** is a

The
solution to such an instance is a sequence ** J_{1}, …, J_{n}** determining the order in which the
jobs are executed, and a sequence

Processing
each job requires a particular collection of tools loaded in the magazine. It
is assumed that no job requires a number of tools higher than the magazine
capacity, i.e., for all ** j**, where is Kronecker's delta. The objective function

**Datasets**

As far as
we know, no standard data instance exists for this problem (at least publicly
available) so that we have arbitrarily
selected a wide set of problem instances that were attacked in (Bard, 1988)(Hertz,1998)(AlFawzan,2003)(Zhou,2005);
more specifically, 16 instances were chosen with values for the number of jobs,
number of tools, and machine capacity ranging in [10,50], [9,60] and [4,25]
respectively. Table 1 shows the different problem instances chosen for the
experimental evaluation where a specific instance with ** n** jobs,

*Table 1.** Problem Instances considered in the
experimental evaluation. The minimum and maximum of tools required for
all the jobs is indicated in second and third rows respectively. Fourth row
display the bibliography reference from which the problem instance was obtained.
[1](Bard, 1988)[13](Hertz,1998)[19](AlFawzan,2003)[20](Zhou,2005).*

Five
different datasets (i.e., incident matrixes or relations among tools and jobs)
were generated randomly per instance. Each dataset was generated with the
restriction, already imposed in previous works such as (Hertz, 1998), that no
job is ** covered** by any other job in the sense that where is defined
as before, i.e., the set of tools required to process job

*Format of datasets*

Each file has
follow structure.

The file is
composing by** n** lines. Each line have follow information

0:1,6,0,4#

this
indicate that job 0 required tools 1,6,0 and 4.

Additionally,
the name file indicates the problem instances used, e.g., matrix_** n**j_

Last update:
17 July 2008 Comments:
jedgar@unet.edu.ve