TASK ALLOCATION AND PATH PLANNING FOR MULTI AUTONOMOUS FORKLIFT NAVIGATION WITH COMBINATION OF HUNGARIAN METHOD AND CONFLICT PRIORITY-BASED SEARCH

• Main Author: Kurnianto, Hardi
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/73244
• Institution: Institut Teknologi Bandung

The development of the e-commerce business affects the demand and volume of products distributed by the warehouse system and causes piles of task queues. The presence of an autonomous multi-forklift system is expected to be a solution. However, the multi-autonomous forklift system has navigation problems that require it to deliver products without conflict between autonomous forklifts. In addition, it must be able to fulfill performance functions such as makespan which is the maximum number of steps that must be taken and the minimum total path cost. This study proposes a modified A-star algorithm pathfinding method (Improved A-star) with the addition of a heading angle variable in the heuristic function to reduce the number of inflection points in the path analysis process. The task allocation process with the Hungarian algorithm which has been modified in the data preprocessing process for assignments and solutions from variations in the number of tasks and the number of available forklifts. The Conflict Priority-based Search (CPBS) algorithm is a modification of the Conflict-based Search (CBS) algorithm with the addition of priority rules for resolving conflicts and stopping solutions for forklifts with low priority. Based on the simulation results that have been carried out, the average difference in inflection point produced by the Improved A-star algorithm is 41.2% better than the original A-star algorithm. Then the use of the modified Hungarian algorithm can allocate tasks to a variety of quantities well. When compared with Market-based Auction. The results for the same number of tasks and the number of forklifts are better with an average of 3.7% in total line costs and 7.26% in makespan. In the condition that the number of forklifts is more, the average is better, 6.1% in total line costs and 3.95% in makespan. If the number of tasks is higher, the average is 4.87% better in line costs and 6.8% in makespan. Then the search and conflict resolution using the CPBS algorithm has a better success rate on the number of forklifts of more than 6 forklifts with a shorter total path cost of 60.41% even though it has a 64.83% larger makespan..