The randomized time-varying knapsack problem (RTVKP) is both a kind of dynamic knapsack problem and a kind of dynamic combinational optimization problem. Currently, the leading algorithms for solving RTVKP include the exact algorithm base on dynamic programming, approximation algorithm base on greedy-choice strategy and evolutionary algorithm base on genetic algorithm. First, in this paper, an exact algorithm base on dynamic programming to solve RTVKP is presented, along with comparison of its time complexity with the existing exact algorithms. Results show that the proposed algorithm is more suitable to solve RTVKP whose profit is larger. Then, the greedy correction and optimization strategy is combined with differential evolution and particle swarm optimization respectively to solve RTVKP. The numerical results on 5 instances of RTVKP show that the evolutionary algorithms which combine the differential evolution, particle swarm optimization and genetic algorithm with Greedy correction and optimization strategy respectively are more suitable to solve the hard RTVKP whose scale and oscillation frequency are larger while having bigger data.