| Currently,in the context of digital transformation and upgrading,path planning in logistics and transportation and flexible job-shop scheduling in workshop production are attracting attention as key problems affecting digital transformation and upgrading,and the exponential level of data and complex scenarios in industrial production have invariably increased the difficulty of their solutions.A large number of scholars have abstracted the two types of problems into combinatorial optimization problems by modeling them,and studied and designed various algorithms to pursue the optimization of the problems.The existing mainstream algorithms for solving combinatorial optimization problems are exact methods,heuristics,metaheuristics,neural networks,and reinforcement learning for various aspects such as computational accuracy,computational time,and computational cost.As a classical representative of metaheuristic methods,the ant colony optimization algorithm has shown excellent results in solving various combinatorial optimization problems by using its evolutionary search process and evolutionary learning idea,its positive feedback capability and powerful robustness property.However,there are still some shortcomings of the ant colony algorithm,such as the tendency to fall into local optimum and the lack of convergence ability in the late stage.In this paper,we study and design a series of ant colony optimization algorithms to solve combinatorial optimization problems from the perspectives of improving the algorithm structure,introducing heuristics and integrating various strategies.The main work of this paper is summarized as follows.(1)A heterogeneous adaptive based ant colony optimization algorithm is proposed.Based on the classical Maximum-Minimum ant algorithm,this paper isomerizes the traditional population,and further strengthens the influence of individuals in the population by taking Lévy flight for parameter variation;in the process of evolution,the elite individuals with good performance will guide the whole population so as to maintain the efficiency of search;in the path construction stage,the pheromone reconstruction mechanism is designed to maximize the avoidance of the situation of being trapped in local optima for a long time.In the test case of small-scale travelling salesman problem,the convergence speed of the proposed algorithm is greatly improved,the average solution error can be controlled within 2%,and the best stability performance is also achieved.(2)A smoothing heuristic-based ant colony optimization algorithm is proposed.In order to deal with the problems of long computational resource consumption caused by the process of searching for optimal solutions of combinatorial optimization problems by various optimization algorithms,this paper uses an improved heuristic method as the main technical means.The diversity of the population is reasonably preserved by exploiting the adaptive nature of heterogeneous populations,and the individuals with good performance dominate the heterogeneous population;relying on the high efficiency of heuristic method search,three heuristics are designed to compare and validate the important guiding role of smoothing techniques for ant colony optimization algorithms;the iterative optimal solutions obtained during the evolutionary process are used to filter the effective difference information,and the pheromone update mechanism based on difference processing is redesigned to obtain more high-quality solution sets with less computational resources.Two evaluation metrics are used to help the algorithm complete evolutionary state assessment and adjustment,thus guiding the algorithm to be able to switch between mining and exploration states in a timely manner.In the test example of medium and large-scale travel quotient problem,the proposed algorithm has excellent performance in the three aspects of search speed,convergence accuracy,and algorithm stability improved by 1-2 orders of magnitude,and the comprehensive performance also ranked first.(3)Research and application for flexible job shop scheduling problem.Based on the above design of ant colony optimization algorithm,this paper conducts modeling around flexible job shop scheduling problem,effectively codes the workpieces,processes and equipment,etc.,completes the performance verification.Aiming at the real data set collected for flexible job shop scheduling problem in local furniture industry,the proper scheduling model is established first,and the minimum solution of the maximum completion time is completed,and the optimization application of the proposed algorithm in industrial problems is preliminarily realized.The two ant colony optimization algorithms proposed in this paper have completed simulation experiments in travelling salesman problem and flexible job-shop scheduling benchmark test sets and experimental validation in a local furniture industry workshop scheduling problem.Therefore,the proposed strategies and improved algorithms in this paper have some application prospects while completing the theoretical problem solution. |
PDF Full Text Request