| In recent years,environmental pollution and energy consumption have attracted more and more attention,and the traditional manufacturing industry is facing new challenges.The energy consumed by manufacturing industry accounts for about 70% of industrial energy consumption,so there is a large space for energy conservation.Rational allocation of resources and reduction of energy consumption can help enterprises improve efficiency,reduce production costs and give enterprises greater competitiveness.As an extension of job shop scheduling,hybrid flow shop can meet the requirements of personalized and diversified production.It is a hot issue of job shop scheduling in recent years.Therefore,this thesis studies the energy consumption optimization strategy and method of hybrid flow shop problem HFSP(Hybrid Flow Shop Problem).The main contents of this thesis are as follows:(1)Firstly,since HFSP is essentially a combination problem of single machine scheduling and parallel machine scheduling,it can be decomposed into two subproblems:equipment selection and workpiece scheduling,and the corresponding scheduling rules Mar(Machine Allocation Rule)and JSR(Job Sequence Rule)are established respectively.Then the two are combined to produce the combined scheduling rules for solving HFSP.(2)Secondly,in order to reduce the energy consumption of HFSP,considering the actual production constraints of machine startup and shutdown and different machine speeds,the workshop energy consumption is analyzed.The total energy consumption can be divided into four parts: machine startup and shutdown energy consumption,processing energy consumption,idle energy consumption and public energy consumption.A scheduling model aiming at minimizing the sum of weighted delay cost and total energy consumption cost is constructed.At the same time,the time of use price mechanism is introduced to reasonably arrange production activities and try to avoid arranging processing activities at the peak of power consumption,so as to reduce energy consumption.(3)Then,Parallel Gene Expression Programming(PGEP)is proposed to automatically construct efficient scheduling rules to solve HFSP.PGEP designs parallel computing,distributed computing and individual migration strategy based on penetration theory on the basis of GEP,which improves the computing speed and solution quality of GEP method.(4)Finally,the performance of scheduling rules generated by PGEP is evaluated through the simulation experiment of example data.In order to verify the effectiveness of PGEP,the following experiments are set up and compared: 1)PGEP is compared with Hybrid Particle Swarm Optimization(HPSO),a heuristic algorithm with tabu search.The results show that the scheduling rules produced by PGEP algorithm have better performance;2)Comparing PGEP with serial GEP,the results show that parallel and distributed computing can significantly improve the solution speed of PGEP,and the migration between populations can increase the diversity of populations and produce better scheduling schemes;3)Through the comparison between PGEP in offline training and online solving mode and PGEP in direct solving mode,it is found that the scheduling rules constructed by PGEP after learning and training are better than those generated by PGEP in direct solving instance. |
