A novel harmony search algorithm based on teaching-learning strategies for 0-1 knapsack problems.

Area of Science:

• Optimization Algorithms
• Computational Intelligence
• Operations Research

Background:

• Discrete optimization problems, such as the 0-1 knapsack problem, present significant computational challenges.
• Existing harmony search (HS) algorithms require enhancements for effective discrete problem-solving.
• Balancing global exploration and local exploitation is crucial for algorithm efficiency.

Purpose of the Study:

• To propose a novel harmony search algorithm (HSTL) integrating teaching-learning strategies for 0-1 knapsack problems.
• To enhance the performance of the harmony search algorithm in discrete optimization contexts.
• To dynamically adjust algorithm parameters for improved exploration-exploitation balance.

Main Methods:

• A novel harmony search algorithm based on teaching-learning (HSTL) strategies was developed.
• Dynamic dimension adjustment for harmony vectors was implemented.
• Four key strategies were integrated: harmony memory consideration, teaching-learning, local pitch adjusting, and random mutation.
• Dynamic parameter adjustment strategies were employed to balance exploration and exploitation.

Main Results:

• Simulation experiments were conducted on 13 benchmark 0-1 knapsack problems.
• The HSTL algorithm demonstrated superior performance compared to existing methods.
• The dynamic parameter adjustment effectively maintained a balance between global exploration and local exploitation.

Conclusions:

• The proposed HSTL algorithm is an efficient and effective approach for solving 0-1 knapsack problems.
• Dynamic strategy integration and parameter adjustment significantly enhance HS algorithm performance.
• HSTL offers a promising alternative for discrete optimization tasks.