Results And Conclusions:Future Scope

Initially we have tried with the number of population is twice of the number of the machines which result the earlier convergence of the problem. Due to this there may be possible of stuck at local optima. So we fixed up the number of population to 30. It gives better result than the former.
For the further improvement, we gave attention to the selection parameter.
In the beginning we use proportionate selection method, in general it is used for maximize problem. Since the JSSP is the minimization problem, the linear ranking method is used which is robust and in with the actual fitness (elapse time) is converted into subjective fitness and then the actual counts of selection of each chromosomes was determined.
In continual improvement, we concentrated on the crossover, at first half above the length of the string only consider for the crossover as an initial try. Then the crossover was performed for entire length which results better solutions than former one.
Thus we conclude that there may be a possibility of better solutions if the others operational parameter of GA is used in place of above said methods because One of the key elements for the success of GA is the choice of various parameters values such as population size, number of generation, fitness function, crossover rate and mutation rate and these parameters are generally interact with each other in non linear manner and as a result they cannot be optimized one at a time. Also they may differ for different types of problems and so there are no conclusive results as to what values should be chosen.

Future scope, Job side sequencing with same order of machines may be extended to the different order of machines and also simultaneously machine side sequencing can be performed to get better solutions and also to have different sequence of jobs for each machines.©

For the specific problem as given below the output at the time of execution of problem is shown below.
Output for execution of specific problem
 
For one above sequence the elapse time table is as shown below. In general concept any problem may have many optimal solutions as above.
Elapse time table of specific problem

Related Posts:

• ABSTRACT
• INTRODUCTION
• History
• Biological Background
• Implementation
• Our Approach

Table of Contents
© 2006 Kumaravel & Project Team