REMS: a unified solution representation, problem modeling and metaheuristic algorithm design for general combinatorial optimization problems

Researchers introduce REMS, a unified framework for solving combinatorial optimization problems that views problems as resource allocation tasks. The framework enables reusable metaheuristic algorithms and outperforms established solvers like GUROBI and SCIP on large-scale instances across 10 different problem types.

• →REMS provides a unified paradigm for modeling diverse combinatorial optimization problems as resource-task assignments.
• →The framework eliminates the need for handcrafted algorithms for each specific optimization problem.
• →Five metaheuristic algorithms were developed using the unified solution structure and fundamental operators.
• →Testing on 10 different problem types showed REMS outperforms GUROBI, SCIP, and OR-TOOLS on complex instances.
• →The approach covers routing, location, loading, assignment, scheduling, and graph coloring optimization problems.