Faster space-efficient algorithms for subset sum and k-sum
Nikhil Bansal, Shashwat Garg, Jesper Nederlof, Nikhil Vyas · 2017
We present randomized algorithms that solve Subset Sum and Knapsack instances with n items in O*(20.86n) time, where the O*(·) notation suppresses factors polynomial in the input size, and polynomial space, assuming random read-only access to exponentially many random bits. These results can be extended to solve Binary Linear Programming on n variables with few constraints in a similar running time. We also show that for any constant k≥ 2, random instances of k-Sum can be solved using O(nk-0.5(n)) time and O(logn) space, without the assumption of random access to random bits.