Lab Assessment 2 & Practical Session 8

Order Description

What you need to submit
Please submit the following files via the coursework submission dropbox link on unit MOODLE page:
1. Soft copy of your report on Moodle. Your report should be no longer than the specified word limits for each part. The font size used must be 10 or greater.
2. Your Excel files on Moodle. 3. You need to submit each file (Word / Excel file) to the corresponding part (tab) within the dropbox on Moodle.

Flow Shop Scheduling problem
A car manufacturing company believes that the demand for a special family car will increase in the following years. So, they want to expand their production capacity by efficiently allocating jobs to machines in the shop floor. In one of their main tasks, the company needs to schedule the processing of 50 components in 5 workstations so as to minimise makespan (the completion time of the last job on the last machine).
Each component is processed by a certain workstation for a certain processing time. The components must be processed in the same sequence in all workstations.
Each component can be processed in only one workstation at a time, and each workstation can process only one component at a time. Operations are not preemptable and the processing times of the components in the workstations are provided in the file “Raw Data – Lab Assessment 2” on MOODLE. The objective is to determine a schedule to process 50 components in 5 workstations that minimises the makespan.

1. Formulate the mathematical optimisation model for the mentioned flow shop scheduling problem to minimise the makespan in the report. [Maximum 200 words]

2. Implement the mathematical model on Evolver and then solve it by experimenting with the following crossover and mutation rates.
i. Crossover rate = 0.9, mutation rate = 0.01.
ii. Crossover rate = 0.8, mutation rate = 0.025.
iii. Crossover rate = 0.6, mutation rate = 0.2.
iv. Crossover rate = 0.4, mutation rate = 0.005.
v. Any other combinations that you think will produce a good solution.

The stopping criterion for all the experiments should be set to 20,000 generations. In the report, provide a detailed table of the best solution of each combination including: sequence of jobs, makespan, and the computation time required to generate the best solution. [Maximum 200 words]

3. Describe the Evolver model and explain clearly the implementation of the decision variables, fitness function and constraints in the report. A number of screenshots of the Excel spreadsheet and Evolver model should be included in the description. [Maximum 500 words]

4. Solve the problem by choosing OptQuest as the optimisation method. Calculate the % deviation of the best minimum makespan value found for each combination in question 2 from the OptQuest solution. Present the % deviation calculations and the results in the report. [Maximum 200 words]

5. Apply Tabu Search (TS) to solve this problem considering only 10 components and 5 machines and explain in detail the TS pseudo-code in the report. Provide an example of the feasible initial solution, neighbourhood structure, tabu list structure, and tabu moves. Run manually TS for 2 iterations to solve the problem. For each iteration, describe in detail the moves used, the solution, the tabu list, and the tabu moves. Provide the best mekaspan found after 2 iterations. [Maximum 500 words]