A Hybrid Approach Combining Improved Shuffled Frog-Leaping Algorithm With Dynamic Programming for Disassembly Process Planning

Hsien-Pin Hsu,Chia-Nan Wang

doi:10.1109/access.2021.3072831

Abstract

This paper proposes a hybrid approach which combines Improved Shuffled Frog-Leaping Algorithm (ISFLA) with Dynamic Programming (DP) for disassembly process planning (DPP). The DPP is a new frontier to the end-of-life product management which can help protect our environment. In the DPP, two tasks are to be accomplished. One is the determination of a disassembly sequence for components and another is the assignment of each component with a responsible operator. In the hybrid approach, the Shuffled Frog-Leaping Algorithm is used to accomplish the first task while the dynamic programming is used to complete the second task. Preliminary experimental results show the proposed hybrid approach outperforms the standard Shuffled Frog-Leaping Algorithm and improved Shuffled Frog-Leaping Algorithm. The comparison to a MILP has further confirmed that the proposed hybrid approach can find the optimal/near-optimal solution in terms of disassembly time. Further experiments show that the hybrid approach outperforms a genetic GA and MA. The results show the potential of using hybrid approaches for disassembly process planning.

Highlights

Economic development has introduced various products into our environment
The results show that the hybrid approach performs the best as it has an average gap of 5.2% over the improved Shuffled Frog-Leaping Algorithm (SFLA) (ISFLA) and an average gap of 48% over the SFLA. 4) Table 6 shows the results obtained from different approaches at the problem size 80 × 80
In this research, the experimental results show that the memetic algorithm (MA) is still inferior to the proposed Hybrid approach (ISFLA + Dynamic Programming (DP)) when dealing with the disassembly process planning (DPP). 12) Algorithm 4 shows the outline of the Hybrid approach (ISFLA + DP)

Summary

INTRODUCTION

Economic development has introduced various products into our environment. One of main kinds is electronic products such as personal computers, telephones, televisions, refrigerators, etc. One kind of the meta-heuristics is natureinspired and swarm intelligence (SI) based meta-heuristics such as Genetic Algorithm (GA), Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), Bee-inspired Algorithms (BA), Bacterial Foraging Optimization (BFO), Firefly Algorithm (FA), Fish Swarm Optimization (FWO), and Shuffled Frog-Leaping Algorithm (SFLA) These metaheuristics have gained increasing attention due to their unique advantages such as the improvement of simplicity of general heuristics and the prevention of computational intractability of exact approaches. B. DISASSEMBLY UNCERTAINTIES CONSIDERED AND THE MEMBERSHIP FUNCTIONS USED Referring to [16], the (1) shows the time required for operator j to disassemble component i with the skill of the operator and product condition taking into account. The R is a random number within [0,1]; Xb(t) is the position of the best frog in the sub-memeplex; Sm is the maximum leaping step allowed.

THE ISFLA The ISFLA has the following novel features

NUMERICAL EXPERIMENTS

DISCUSSION AND ANALYSIS Discussion and findings are given as follows

Initial positions for frogs

Findings

CONCLUSION