Page 225 - Mechatronics for Safety, Security and Dependability in a New Era
P. 225

Ch43-I044963.fm  Page 209  Tuesday, August 1, 2006  3:58 PM
                                           3:58 PM
                            Tuesday, August
                      Page 209
                                      1, 2006
            Ch43-I044963.fm
                                                                                          209
                                                                                          209
                  REACTIVE SCHEDULING METHOD FOR TARDINESS MINIMIZATION       PROBLEMS
                  Tardiness minimization  problems
                  In  this  paper,  the  main  concern  is total tardiness, which  is  a  criterion  based  on job  due  dates,  and  is
                  defined  by Eqn.  1.

                                                 £max(0,C,. -d t)                         (1)
                                                 i=\

                  Where  Q  and d t are the  completion  time  and the  due  date  of the job J t respectively,  and n is the total
                  number of the jobs under  consideration.
                  The  reactive  scheduling  process  using  GA  is  a time-consuming  process  in  order  to  generate  a  good
                  solution. However, the reactive  scheduling modifies  the  schedule  in parallel to the production  activity,
                  therefore,  it  is required  to  find  a good  solution  in the  limited  time. An  advanced  crossover  method  is
                  discussed in the followings,  in order to modify  and improve the schedule in a short time.

                  Advanced  crossover  method
                  Each gene  in an individual  of the proposed  GA method corresponds to a manufacturing  operation to be
                  executed  in  the  manufacturing  system,  and  the  list  of  the  genes  in  the  individual  represents  the
                  priorities  for  the  execution  of manufacturing  operations  in the production  schedule.  The  lower  bound
                  of the tardiness is estimated  for the _y-th gene of the individual, by applying Eqn. 2.


                                                               dd,                        (2)

                  where,
                  LTy. lower bound of tardiness  for the >"-th gene, which corresponds to the h-th operation of job Jj.
                  ft,:  finishing time of the h-th operation of job </,.
                  pt f^  (s=h+\, ...,«): processing time of remaining operations of job Jj.
                  dd (. due-date of job J,-.

                  If  the LT y  is  more  than  zero,  it  is  impossible  for job J t  to  finish  the  remaining  operations  by  its  due
                  date.

                  Two  parent  individuals  and  their  crossover  points  are  randomly  selected  in  the  first  step  of  the
                  crossover  operation.  After  that,  only  the genes  having positive  number  of LT y between  two  crossover
                  points  are  exchanged  with  the  genes  of  another  parent  individual,  by  the  newly  proposed  crossover
                  method.  The  other  genes  of the  parent  individuals  are  survived  to the  offspring  individuals,  as  shown
                  in Figure 2.

                  Reactive scheduling  process
                  The reactive scheduling process  is carried out by the following  steps.

                  STEP1 Initialization
                  The present time T x (x 1,2,...)  is set up. Computation time dt is estimated. It is the time  for  creating
                                  =
                  the modified  production  schedules through  STEP2 to STEP4.
   220   221   222   223   224   225   226   227   228   229   230