Research On Shop Scheduling Problems Based On Analytical Target Cascading And Intelligent Algorithms

Production scheduling allocates the shared resources available and arranges processingtasks in a period of time, to meet one or some of the specific production targets. The researchand application of effective scheduling methods and optimization technology are thefoundation and key to promote advanced manufacturing and production efficiency. Optimalproduction scheduling schemes can significantly improve production efficiency and resourceutilization, and further to enhance enterprise ability to compete, so the production schedulingproblems have been a hot topic in manufacturing systems. However, job shop scheduling is atypical combinatorial optimization problem, which is a typical NP-hard problem, so theproblem is one of the most difficult problems in theory. At the same time, with thedevelopment of networked manufacturing, there are more and more optimization problems fordistributed multi-shop scheduling; the remote distribution of production resources andproduction facilities located in different factories or workshops, and the distribution andnon-predictability of manufacturing information make the existing production schedulingsystem, method and strategy difficult to produce the desired results. To solve such problems,distributed job-shop scheduling based on analytical target cascading and intelligent algorithmsare investigated as follows:An entropy-enhanced genetic-based tabu search algorithm is proposed and appliedin a single workshop. After the basic model of job-shop scheduling problem described, thebasic process of genetic algorithm and control parameters are introduced, and encoding anddecoding, selection, crossover, mutation and fitness function of the genetic algorithm areinvestigated for workshop scheduling problems. Combining the tabu search algorithm, geneticalgorithms and population entropy, a so-called entropy-enhanced genetic-based tabu searchalgorithm is proposed and applied to solve job shop scheduling problems.An entropy-enhanced Particle Swarm Optimization (PSO) algorithm is proposedand applied in a single workshop. After the basic models for flexible shop schedulingproblems introduced, the basic principle and characteristics of the PSO algorithm aredescribed, and the PSO algorithm is applied in job shop scheduling problems, and especiallythe design of location vector, position vector and velocity vector areemphasized. According tothe characteristics of PSO, simulated annealing and population entropy, a so calledentropy-enhanced PSO is proposed, and applied to solve a single flexible shop schedulingproblem.The hybrid intelligent optimization algorithm is applied in a distributedhierarchical shop scheduling problem. For distributed shop scheduling problems, a two -layer scheduling optimization model is established, whose production planning layer isresponsible for the distribution of part orders, and shop scheduling layer plans the processingpath of parts. The proposed hybrid genetic algorithm is used to solve the distributed shopscheduling problems, and the proposed hybrid particle swarm algorithm used to solvedistributed flexible job shop scheduling problems.Analytical target cascading is applied in large-scale shop scheduling and flexibledistributed shop scheduling problem. The model of a large-scale job shop schedulingproblem is established based on the analytical target cascading, which includes the model ofthe part family planning layer and that of part path planning layer, and is verified withexamples. And analytical target cascading is applied in flexible distributed shop schedulingproblems, and validated with examples, too.Particle swarm optimization algorithm and Analytical target cascading are appliedin solving multi-objective flexible job shop scheduling problems. Particle swarmoptimization is applied in a single-shop scheduling problem with the objectives of completetime, machining cost and delivery, and the multi-objective distributed flexible job shopscheduling model is established based on analytical target cascading, which includes themodel of the production planning layer and that of the shop floor scheduling layer, andverified with examples.Scheduling software design and implementation. First the software developmentenvironment for scheduling is analyzed, which includes the choice of development platformsand databases, and then described are software design and implementation for themanagement layer, production scheduling layer and shop scheduling layer. The shopscheduling with genetic algorithm for FT06and the distributed shop scheduling based ontarget cascading are taken as examples to verify the effectiveness of the developed schedulingsoftware.