Study On Resources-Constrained Project Scheduling Problem On Improved Cuckoo Algorithm

In recent years,with the rapid development of information technology such as intelligent computing technology,mobile network communication,and related disciplines,the new generation of information intelligence represented by big data analysis and cloud computing systems is being integrated into people's daily lives.From the perspective of resources,big data and cloud computing technology are a new type of resource,which embodies a brand-new view of resources.In the actual project management problem,project scheduling is the most important link,and resource constraints also greatly affect the project scheduling plan.Resource-constrained project scheduling problem(RCPSP)is a classic project scheduling problem model,which is widely used in various fields such as cloud computing resource scheduling,communication broadband allocation,and logistics scheduling.Many classic scheduling problems,such as assembly line problems,job shop problems,etc.,are based on the transformation of the RCPSP problem,so the research on the RCPSP problem has important academic significance and practical value.The subject takes the RCPSP problem as the main research object.Aiming at the problems and defects of the basic cuckoo algorithm,an improved cuckoo algorithm(NCS&SA)is proposed.First,the random key vector coding scheme is adopted to encode individuals in the population as priority vectors of random keys,and the priority vectors are sorted,so that the relationship between priority and index is transformed into the relationship between task and index.Through the serial scheduling mechanism,the encoded individuals of the population are decoded into the initial task scheduling sequence;secondly,the basic cuckoo algorithm has great randomness when searching for Levy flights,which affects the search efficiency and quality of the algorithm.Therefore,this paper changes the probability density function of Levy flights to be adaptive as the population fitness changes.Then three neighborhood update technologies,neighborhood exchange,neighborhood insertion,and 2-opt are added.According to the probability density calculated by Levy flights,the corresponding neighborhood update technologies are selected at different stages of the algorithm to enhance the ability of the algorithm to locally optimize.Finally,in order to prevent the algorithm from falling into the local optimum,this paper introduces the simulated annealing strategy to accept individuals with poor quality with a certain probability in each iteration of the algorithm,and enhance the algorithm's global optimization ability and convergence accuracy.In the algorithm test part,this paper selects the classic example library PSPLIB for RCPSP problems to test the convergence and optimization of the algorithm.The average error of the NCS&SA algorithm on the small-scale data of J30 is 0.25%,and the average error of the large-scale data of J60 is 11.21%.The average error of J120large-scale data is 19.83%.Experiments prove that the convergence,convergence accuracy and optimization of the algorithm are better than the basic cuckoo algorithm and other intelligent algorithms.In order to test the ability of the NCS&SA algorithm in practical engineering applications,this paper selects a practical engineering scheduling case for test analysis.The results prove that the NCS&SA algorithm is superior to the basic cuckoo algorithm in solving practical problems and has practical engineering application capabilities.