Research On Multi-objective Scheduling Of Green Shop Based On Swarm Intelligence Algorithm

Shop scheduling which is an important part of manufacturing system,realizes energy saving,emission reduction,quality improvement and efficiency by optimizing resource allocation,task assignment and operation sequencing.Compared with the traditional shop scheduling,the multi-objective shop scheduling problem has the characteristics of multiobjective,complex constraints,high dimensions,large scale,nonlinear,etc.,making it more difficult to solve,and its modeling and optimization are extremely challenging.Swarm intelligence algorithm shows good performance in multi-objective scheduling problem of shop.This kind of method is convenient to use and easy to understand.When it is applied to solve shop scheduling problem,the mathematical properties of the problem solved are not strictly required,and it is easy to obtain satisfactory scheduling solutions.Swarm intelligence optimization algorithm provides a new and effective method to solve the multiobjective scheduling problem,which is the frontier research hotspot in the field of intelligent manufacturing academia.As the largest manufacturing country in the world,the manufacturing industry promotes the rapid development of China,but also brings a series of serious consequences such as high energy consumption,high emission and high pollution.It has important theoretical value and practical significance to study the multi-objective scheduling problem of workshop.In this context,this dissertation designs several improved swarm intelligence optimization multi-objective algorithms to solve the problems,such as parallel machine scheduling;hybrid flow shop scheduling;flexible job shop scheduling.The core contents are summarized as follows:For the parallel machine scheduling problem,this dissertation develops the mathematical programming model with objectives of minimizing total tardiness penalty cost and noise pollution.As for solution generation,a multiple-objective shuffled frog leaping algorithm(MOSFLA)is proposed by virtue of the shuffled frog leaping algorithm(SFLA)and Pareto dominance concept.In the algorithm design,a novel two-layer encoding schema is presented according to the problem characteristics.Meanwhile,the population during the iteration is sorted by the population stratification and crowding distance values.Moreover,different crossover operators(e.g.,PBX,CX and SEX)are employed to define the population evolution method,and its efficiency is validated by the controlled experiments.Comprehensive results between MOSFLA and other MOEAs under different instance scales verify its outstanding performance in solving the proposed parallel machine scheduling problem.For the hybrid flow shop scheduling problem,a mathematical model is established on purposed of optimizing makespan and noise pollution.A novel MOEA named MMOSFLA(modified multiple-objective shuffled frog leaping algorithm)is developed on the basis of the proposed MOSFLA.In the algorithm design,the priority sequence and velocity matrix are embedded into solution representation according to the problem nature.Furthermore,a decomposition-based archive maintenance is introduced to enhance MMOSFLA's performance in aspects of diversity and convergence.Experiments are presented to compare the proposed scheduling method with other classical MOEAs.Comprehensive results demonstrate that the proposed MODGWO is significantly better than benchmark algorithms.For the flexible job shop scheduling problem,this dissertation establishes a mathematical programming model on purpose of minimizing makespan and noise pollution.A modified multiple-objective artificial bee colony(MMOABC)algorithm is established based on the proposed MODABC.MMOABC adopts a novel three-layer encoding method to represent individuals in the algorithm,and an active scheduling decoding mechanism is utilized to obtain high-quality solutions.In addition,a filter-and-fan technique is employed to enhance the algorithm's exploitation ability by virtue of the critical chain concept.Experimental results demonstrate that MMOABC is superior to benchmark MOEAs on most instances.Two case studies are provided to investigate scheduling problems encountered in a welding workshop and an engine cooling fan production system,respectively.The above theoretical fruits are applied to formulate the detailed problems and to generate schedule solutions.Experimental results validate the effectiveness of the proposed multi-objective scheduling methods.Finally,the innovative points of the above research work are summarized,and the future research directions are discussed.