Research On Distributed Hybrid Flow-shop Scheduling Method Based On Artificial Bee Colony Algorithm

Distributed manufacturing has become a very popular topic in the field of manufacturing,of which distributed shop scheduling is the difficult part of distributed manufacturing.Distributed shop scheduling splits the production tasks into multiple sub-tasks,which are scheduled and completed in different shops.This results in efficient production.Distributed Hybrid Flow-shop Scheduling Problem(DHFSP)is a complex distributed shop scheduling problem that efficiently coordinates the resource allocation of multiple hybrid flow-shops by distributing jobs between shops and sorting and selecting machines within shops.It is widely used in many fields such as mechanical engineering,medical services,logistics and warehousing.This work takes DHFSP as the research object,constructs a mathematical model of the problem,and uses the Artificial Bee Colony(ABC)algorithm to study in depth the extended problems of homogeneous shop environment,heterogeneous shop environment and multi-objective optimisation considering multiple types of constraints,and validates the application with practical engineering.The main work are shown below:First,an ABC algorithm is proposed for the classical hybrid flow shop scheduling problem.For the encoding and decoding method,combined with the representation of the parsing graph,a critical path search method based on hybrid bi-directional scheduling is proposed and introduced into the ABC algorithm,which solves the problem of poor search performance in the sequence encoding space;in the onlooker bee stage,a tournament selection mechanism without put-back is designed to ensure the diversity of the population and the quality of solutions;in the scout bee stage,an elite strategy of inversion operation is constructed that improves the breadth of the search.Experimental results on three standard test sets validate the effectiveness of the improved ABC algorithm and proposed strategies.Secondly,a mathematical model with the objective of minimising the maximum completion time and a improved ABC algorithm are proposed for the homogeneous DHFSP.Based on the solution space of the problem,a two-level encoding and a decoding method based on three shop assignment rules are proposed to achieve a hybrid scheduling for multishop characteristics.A neighbourhood search strategy based on key shop and key paths is designed to improve the depth of the algorithm search.Based on the characteristics of the objective function,a fast evaluation method is proposed to improve the search efficiency of the algorithm.Comparative experiments verify the effectiveness of the improved ABC algorithm and proposed strategies.Then,based on a homogeneous DHFSP,a heterogeneous DHFSP is proposed,taking into account the differences in the processing environment of each shop and the Sequence Dependent Setup Time(SDST)constraint,and a mathematical model and an improved ABC algorithm are constructed with the objective of minimising the maximum completion time.A hybrid scheduling method for SDST and assignment rules based on heterogeneous shops are proposed to achieve an efficient search of the sequence encoding.For the flexible characteristics of machines,greedy search on critical paths of critical shops is proposed,and the quality of the neighbourhood solution is improved as well.Comparative experiments verify the effectiveness of the ABC algorithm and the superiority of improved strategies.Then,based on heterogeneous DHFSP with SDST,a multi-objective heterogeneous DHFSP with multiple constraints is proposed by considering constraints of batch processor,lot streaming and transport time.A mathematical model with the objective of minimising the maximum completion time and the number of sub-batches is constructed,and a decomposition-based multi-objective ABC algorithm is proposed.A three-level encoding and a sub-batching strategy are designed to achieve hybrid scheduling.In the employee bee stage,a batch crossover operator and batch adjustment method are proposed to retain the good characteristics of the parent while enhancing the population diversity;in the onlooker bee stage,a clustering move operation is adopted,which can fully explore the encoding space;in the scout bee stage,new solutions are mined based on different weights to achieve the exchange of solutions with the dominant hierarchy;the algorithm enriches the set of nondominated solutions through two update mechanisms of solutions.Numerical experiments comparing the algorithms verify the efficiency of the proposed algorithm and strategies.Subsequently,the theory and methods proposed in this paper are applied to practical engineering case of a domestic mobile phone shell workshop.This case study is reduced to HFSP and homogeneous DHFSP,which are solved using the proposed ABC algorithm in this paper,and verifies the practicality of the proposed algorithm and strategies.Finally,paper summarizes the work of the thesis,and look forwards to the future research direction.