Abstract
Given an n-dimensional convex body by a membership oracle in general, it is known that any polynomial-time deterministic algorithm cannot approximate its volume within ratio \((n/\log n)^n\). There is a substantial progress on randomized approximation such as Markov chain Monte Carlo for a high-dimensional volume, and for many #P-hard problems, while only a few #P-hard problems are known to yield deterministic approximation. Motivated by the problem of deterministically approximating the volume of a \(\mathcal{V}\)-polytope, that is a polytope with a small number of vertices and (possibly) exponentially many facets, this paper investigates the problem of computing the volume of a “knapsack dual polytope,” which is known to be #P-hard due to Khachiyan (1989). We reduce an approximate volume of a knapsack dual polytope to that of the intersection of two cross-polytopes, and give FPTASs for those volume computations. Interestingly, computing the volume of the intersection of two cross-polytopes (i.e., \(L_1\)-balls) is #P-hard, unlike the cases of \(L_{\infty }\)-balls or \(L_2\)-balls.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Precisely, they are concerned with a “well-rounded” convex body, after an affine transformation of a general finite convex body.
- 2.
Given \(\varvec{a} \in \mathbb {Z}_{>0}^n\) and \(b \in \mathbb {Z}_{>0}\), the problem is to compute \(|\{ \varvec{x} \in \{0,1\}^n \mid \sum _{i=1}^n a_i x_i \le b \}|\). Remark that it is computed in polynomial time when all the inputs \(a_i\) (\(i=1,\ldots ,n\)) and b are bounded by \(\mathrm {poly}(n)\), using a version of the standard dynamic programming for knapsack problem (see e.g., [7, 13]). It should be worth noting that [12, 24] needed special techniques, different from ones for optimization problems, to design FPTASs for the counting problem.
- 3.
See [23] for the duality of polytopes. In fact, \(P_{\varvec{a}}\) itself is not the dual of a knapsack polytope in a canonical form, but it is obtained by an affine transformation from a dual of knapsack polytope under some assumptions. Khachiyan [16] says that computing \(\mathrm {Vol}(P_{\varvec{a}})\) ‘is “polar” to determining the volume of the intersection of a cube and a halfspace.’ .
- 4.
If all \(a_i\) (\(i=1,\ldots ,n\)) are bounded by \(\mathrm {poly}(n)\), it is computed in polynomial time, so did the counting knapsack solutions. See also footnote 1 for counting knapsack solutions.
- 5.
Suppose you know that x is approximately 49 within 1% error. Then, you know that \(x+50\) is approximately 99 within 1% error. However, it is difficult to say \(50-x\) is approximately 1. Even when additionally you know that x does not exceed 50, \(50-x\) may be 2, 1, 0.1 or smaller than 0.001, meaning that the approximation ratio is unbounded.
- 6.
We will set \(\beta = 1-\dfrac{\epsilon }{2n\Vert \varvec{a}\Vert _1}\), later.
- 7.
Most of the proofs cannot be included due to the space limit.
- 8.
Remark that \(\mathrm {Vol}(C(\varvec{c},r)\cap C(\varvec{c}',r')) = r^n\mathrm {Vol}\left( C(\varvec{0},1) \cap C\!\left( \tfrac{(\varvec{c}-\varvec{c}')^+}{r}, \tfrac{r'}{r} \right) \right) \) holds for any \(\varvec{c}, \varvec{c}' \in \mathbb {R}^n\) and \(r,r' \in \mathbb {R}_{>0}\), where \((\varvec{c} - \varvec{c}')^+ = (|c_1-c_1'|,|c_2-c_2'|,\ldots ,|c_n-c_n'| )\).
References
Ando, E., Kijima, S.: An FPTAS for the volume computation of 0–1 knapsack polytopes based on approximate convolution. Algorithmica 76(4), 1245–1263 (2016)
Bandyopadhyay, A., Gamarnik, D.: Counting without sampling: asymptotics of the log-partition function for certain statistical physics models. Random Struct. Algorithms 33, 452–479 (2008)
Bárány, I., Füredi, Z.: Computing the volume is difficult. Discrete Comput. Geom. 2, 319–326 (1987)
Bayati, M., Gamarnik, D., Katz, D., Nair, C., Tetali, P.: Simple deterministic approximation algorithms for counting matchings. In: Proceedings of STOC 2007, pp. 122–127 (2007)
Cousins, B., Vempala, S., Bypassing, K.L.S.: Gaussian cooling and an \(O^\ast (n^3)\) volume algorithm. In: Proceedings of STOC 2015, pp. 539–548 (2015)
Dadush, D., Vempala, S.: Near-optimal deterministic algorithms for volume computation via M-ellipsoids. Proc. Natl. Acad. Sci. USA 110(48), 19237–19245 (2013)
Dyer, M.: Approximate counting by dynamic programming. In: Proceedings of STOC 2003, pp. 693–699 (2003)
Dyer, M., Frieze, A.: On the complexity of computing the volume of a polyhedron. SIAM J. Comput. 17(5), 967–974 (1988)
Dyer, M., Frieze, A., Kannan, R.: A random polynomial-time algorithm for approximating the volume of convex bodies. J. Assoc. Comput. Mach. 38(1), 1–17 (1991)
Elekes, G.: A geometric inequality and the complexity of computing volume. Discrete Comput. Geom. 1, 289–292 (1986)
Gamarnik, D., Katz, D.: Correlation decay and deterministic FPTAS for counting list-colorings of a graph. In: Proceedings of SODA 2007, pp. 1245–1254 (2007)
Gopalan, P., Klivans, A., Meka, R.: Polynomial-time approximation schemes for knapsack and related counting problems using branching programs. arXiv:1008.3187v1 (2010)
Gopalan, P., Klivans, A., Meka, R., Štefankovič, D., Vempala, S., Vigoda, E.: An FPTAS for #knapsack and related counting problems. In: Proceedings of FOCS 2011, pp. 817–826 (2011)
Karp, R.: Reducibility among combinatorial problems. In: Miller, R.E., Thatcher, J.W. (eds.) Complexity of Computer Computations, pp. 85–103. Plenum Press, New York (1972)
Khachiyan, L.: The problem of computing the volume of polytopes is \(\#P\)-hard. Uspekhi Mat. Nauk. 44, 199–200 (1989)
Khachiyan, L.: Complexity of polytope volume computation. In: Pach, J. (ed.) New Trends in Discrete and Computational Geometry, pp. 91–101. Springer, Berlin (1993)
Li, L., Lu, P., Yin, Y.: Approximate counting via correlation decay in spin systems. In: Proceedings of SODA 2012, pp. 922–940 (2012)
Li, L., Lu, P., Yin, Y.: Correlation decay up to uniqueness in spin systems. In: Proceedings of SODA 2013, pp. 67–84 (2013)
Li, J., Shi, T.: A fully polynomial-time approximation scheme for approximating a sum of random variables. Oper. Res. Lett. 42, 197–202 (2014)
Lin, C., Liu, J., Lu, P.: A simple FPTAS for counting edge covers. In: Proceedings of SODA 2014, pp. 341–348 (2014)
Lovász, L.: An Algorithmic Theory of Numbers, Graphs and Convexity. Applied Mathematics. SIAM Society for Industrial, Philadelphia (1986)
Lovász, L., Vempala, S.: Simulated annealing in convex bodies and an \(O^\ast (n^4)\) volume algorithm. J. Comput. Syst. Sci. 72, 392–417 (2006)
Matous̆ek, J.: Lectures on Discrete Geometry. Springer, New York (2002)
Štefankovič, D., Vempala, S., Vigoda, E.: A deterministic polynomial-time approximation scheme for counting knapsack solutions. SIAM J. Comput. 41(2), 356–366 (2012)
Weitz, D.: Counting independent sets up to the tree threshold. In: Proceedings of STOC 2006, pp. 140–149 (2006)
Acknowledgments
This work is partly supported by Grant-in-Aid for Scientific Research on Innovative Areas MEXT Japan “Exploring the Limits of Computation (ELC)” (No. 24106008, 24106005) and by JST PRESTO Grant Number JPMJPR16E4, Japan.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this paper
Cite this paper
Ando, E., Kijima, S. (2017). An FPTAS for the Volume of Some \(\mathcal{V}\)-polytopes—It is Hard to Compute the Volume of the Intersection of Two Cross-Polytopes. In: Cao, Y., Chen, J. (eds) Computing and Combinatorics. COCOON 2017. Lecture Notes in Computer Science(), vol 10392. Springer, Cham. https://doi.org/10.1007/978-3-319-62389-4_2
Download citation
DOI: https://doi.org/10.1007/978-3-319-62389-4_2
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-62388-7
Online ISBN: 978-3-319-62389-4
eBook Packages: Computer ScienceComputer Science (R0)