Abstract
In this paper, we develop a simple and fast online algorithm for solving a class of binary integer linear programs (LPs) arisen in general resource allocation problem. The algorithm requires only one single pass through the input data and is free of matrix inversion. It can be viewed as both an approximate algorithm for solving binary integer LPs and a fast algorithm for solving online LP problems. The algorithm is inspired by an equivalent form of the dual problem of the relaxed LP and it essentially performs (one-pass) projected stochastic subgradient descent in the dual space. We analyze the algorithm in two different models, stochastic input and random permutation, with minimal technical assumptions on the input data. The algorithm achieves \(O\left( m \sqrt{n}\right) \) expected regret under the stochastic input model and \(O\left( (m+\log n)\sqrt{n}\right) \) expected regret under the random permutation model, and it achieves \(O(m \sqrt{n})\) expected constraint violation under both models, where n is the number of decision variables and m is the number of constraints. The algorithm enjoys the same performance guarantee when generalized to a multi-dimensional LP setting which covers a wider range of applications. In addition, we employ the notion of the permutational Rademacher complexity and derive regret bounds for two earlier online LP algorithms for comparison. Both algorithms improve the regret bound by a factor of \(\sqrt{m}\) by paying more computational cost. Furthermore, we demonstrate how to convert the possibly infeasible solution to a feasible one through a randomized procedure. Numerical experiments illustrate the general applicability and effectiveness of the algorithms.
Similar content being viewed by others
References
Li, X., Sun, C., Ye, Y.: Simple and fast algorithm for binary integer and online linear programming. In: NeurIPS (2020)
Ferguson, T.S., et al.: Who solved the secretary problem? Stat. Sci. 4(3), 282–289 (1989)
Kellerer, H., Pferschy, U., Pisinger, D.: Knapsack problems. Springer, Berlin, Heidelberg (2004)
Vanderbei, R.J., et al.: Linear Programming. Springer, Boston, MA (2015)
Conforti, M., Cornuéjols, G., Zambelli, G., et al.: Integer Programming, vol. 271. Springer, Cham (2014)
Buchbinder, N., Naor, J.: Online primal-dual algorithms for covering and packing. Math. Oper. Res. 34(2), 270–286 (2009)
Mehta, A., Saberi, A., Vazirani, U., Vazirani, V.: Adwords and generalized on-line matching. In: 46th Annual IEEE Symposium on Foundations of Computer Science (FOCS’05), pp. 264–273 (2005). IEEE
Ford, L.R., Jr., Fulkerson, D.R.: A suggested computation for maximal multi-commodity network flows. Manage. Sci. 5(1), 97–101 (1958)
Dantzig, G.B.: Linear programming and extensions (1963)
De Farias, D.P., Van Roy, B.: On constraint sampling in the linear programming approach to approximate dynamic programming. Math. Oper. Res. 29(3), 462–478 (2004)
Lakshminarayanan, C., Bhatnagar, S., Szepesvári, C.: A linearly relaxed approximate linear program for markov decision processes. IEEE Trans. Autom. Control 63(4), 1185–1191 (2017)
Vu, K., Poirion, P.-L., Liberti, L.: Random projections for linear programming. Math. Oper. Res. 43(4), 1051–1071 (2018)
Calafiore, G., Campi, M.C.: Uncertain convex programs: randomized solutions and confidence levels. Math. Program. 102(1), 25–46 (2005)
Beck, A.: First-order Methods in Optimization, vol. 25. SIAM, Philadelphia, PA (2017)
Yen, I.E.-H., Zhong, K., Hsieh, C.-J., Ravikumar, P.K., Dhillon, I.S.: Sparse linear programming via primal and dual augmented coordinate descent. In: Advances in Neural Information Processing Systems, pp. 2368–2376 (2015)
Wang, S., Shroff, N.: A new alternating direction method for linear programming. In: Advances in Neural Information Processing Systems, pp. 1480–1488 (2017)
Lin, T., Ma, S., Ye, Y., Zhang, S.: An admm-based interior-point method for large-scale linear programming. arXiv preprint arXiv:1805.12344 (2018)
Molinaro, M., Ravi, R.: The geometry of online packing linear programs. Math. Oper. Res. 39(1), 46–59 (2013)
Agrawal, S., Wang, Z., Ye, Y.: A dynamic near-optimal algorithm for online linear programming. Oper. Res. 62(4), 876–890 (2014)
Kesselheim, T., Tönnis, A., Radke, K., Vöcking, B.: Primal beats dual on online packing lps in the random-order model. In: Proceedings of the Forty-sixth Annual ACM Symposium on Theory of Computing, pp. 303–312 (2014). ACM
Gupta, A., Molinaro, M.: How experts can solve lps online. In: European Symposium on Algorithms, pp. 517–529 (2014). Springer
Li, X., Ye, Y.: Online linear programming: Dual convergence, new algorithms, and regret bounds. arXiv preprint arXiv:1909.05499 (2019)
Jasin, S., Kumar, S.: Analysis of deterministic lp-based booking limit and bid price controls for revenue management. Oper. Res. 61(6), 1312–1320 (2013)
Jasin, S.: Performance of an lp-based control for revenue management with unknown demand parameters. Oper. Res. 63(4), 909–915 (2015)
Bumpensanti, P., Wang, H.: A re-solving heuristic with uniformly bounded loss for network revenue management. arXiv preprint arXiv:1802.06192 (2018)
Kleinberg, R.: A multiple-choice secretary algorithm with applications to online auctions. In: Proceedings of the Sixteenth Annual ACM-SIAM Symposium on Discrete Algorithms, pp. 630–631 (2005). Society for Industrial and Applied Mathematics
Arlotto, A., Gurvich, I.: Uniformly bounded regret in the multisecretary problem. Stochastic Systems. 9, 231–260 (2019)
Zhou, Y., Chakrabarty, D., Lukose, R.: Budget constrained bidding in keyword auctions and online knapsack problems. In: International Workshop on Internet and Network Economics, pp. 566–576 (2008). Springer
Balseiro, S.R., Gur, Y.: Learning in repeated auctions with budgets: Regret minimization and equilibrium. Manage. Sci. 65(9), 3952–3968 (2019)
Asadpour, A., Wang, X., Zhang, J.: Online resource allocation with limited flexibility. Manage. Sci. 66, 642–666 (2019)
Jiang, J., Zhang, J.: Online resource allocation with stochastic resource consumption (2019). https://doi.org/10.13140/RG.2.2.27542.09287
Lu, H., Balseiro, S., Mirrokni, V.: Dual mirror descent for online allocation problems. arXiv preprint arXiv:2002.10421 (2020)
Mahdavi, M., Jin, R., Yang, T.: Trading regret for efficiency: online convex optimization with long term constraints. J. Mach. Learn. Res. 13(Sep), 2503–2528 (2012)
Yu, H., Neely, M., Wei, X.: Online convex optimization with stochastic constraints. In: Advances in Neural Information Processing Systems, pp. 1428–1438 (2017)
Yuan, J., Lamperski, A.: Online convex optimization for cumulative constraints. In: Advances in Neural Information Processing Systems, pp. 6137–6146 (2018)
Agrawal, S., Devanur, N.R.: Fast algorithms for online stochastic convex programming. In: Proceedings of the Twenty-sixth Annual ACM-SIAM Symposium on Discrete Algorithms, pp. 1405–1424 (2014). SIAM
Gürbüzbalaban, M., Ozdaglar, A., Parrilo, P.: Why randomreshuffling beats stochastic gradient descent. Math. Program. 186, 49–84 (2021). https://doi.org/10.1007/s10107-019-01440-w
Ying, B., Yuan, K., Sayed, A.H.: Convergence of variance-reduced learning under random reshuffling. In: 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 2286–2290 (2018). IEEE
Safran, I., Shamir, O.: How good is sgd with random shuffling? arXiv preprint arXiv:1908.00045 (2019)
Gao, W., Sun, C., Ye, Y., Ye, Y.: Boosting method in approximately solving linear programming with fast online algorithm. arXiv preprint arXiv:2107.03570 (2021)
Hazan, E.: Introduction to online convex optimization. Foundations and Trends in Optimization 2(3–4), 157–325 (2016)
Devanur, N.R., Hayes, T.P.: The adwords problem: online keyword matching with budgeted bidders under random permutations. In: Proceedings of the 10th ACM Conference on Electronic Commerce, pp. 71–78 (2009). ACM
Panconesi, A., Srinivasan, A.: Randomized distributed edge coloring via an extension of the chernoff-hoeffding bounds. SIAM J. Comput. 26(2), 350–368 (1997)
Tolstikhin, I., Zhivotovskiy, N., Blanchard, G.: Permutational rademacher complexity. In: International Conference on Algorithmic Learning Theory, pp. 209–223 (2015). Springer
Moulines, E., Bach, F.R.: Non-asymptotic analysis of stochastic approximation algorithms for machine learning. In: Advances in Neural Information Processing Systems, pp. 451–459 (2011)
Lacoste-Julien, S., Schmidt, M., Bach, F.: A simpler approach to obtaining an o (1/t) convergence rate for the projected stochastic subgradient method. arXiv preprint arXiv:1212.2002 (2012)
Xiao, L.: Dual averaging methods for regularized stochastic learning and online optimization. J. Mach. Learn. Res. 11(Oct), 2543–2596 (2010)
Juditsky, A., Nesterov, Y.: Deterministic and stochastic primal-dual subgradient algorithms for uniformly convex minimization. Stochastic Systems 4(1), 44–80 (2014)
Flammarion, N., Bach, F.: From averaging to acceleration, there is only a step-size. In: Conference on Learning Theory, pp. 658–695 (2015)
Lei, L., Jordan, M.I.: On the adaptivity of stochastic gradient-based optimization. arXiv preprint arXiv:1904.04480 (2019)
Chu, P.C., Beasley, J.E.: A genetic algorithm for the multidimensional knapsack problem. Journal of heuristics 4(1), 63–86 (1998)
Drake, J.H., Özcan, E., Burke, E.K.: A case study of controlling crossover in a selection hyper-heuristic framework using the multidimensional knapsack problem. Evol. Comput. 24(1), 113–141 (2016)
van der Vaart, J.W.A.: Weak Convergence and Empirical Processes (Springer Series in Statistics). Springer, New York, NY (1996)
Shalev-Shwartz, S., Ben-David, S.: Understanding Machine Learning: From Theory to Algorithms. Cambridge University Press, Cambridge (2014)
Acknowledgements
We thank all seminar participants at Stanford, NYU Stern, Columbia DRO, Chicago Booth, and Imperial College Business School for helpful discussions and comments.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
A preliminary version [1] is accepted in NeurIPS 2020.
Appendices
Appendix
Concentration inequalities under random permutations
Lemma 5
Let \(U_1 , ... , U_n\) be a random sample without replacement from the real numbers \(\{c_1,...,c_N\}\). Then for every \(s > 0\),
where \({\bar{c}}_N = \frac{1}{N}\sum \nolimits _{i=1}^{N}c_i\), \(\sigma ^2_N=\frac{1}{N}\sum \nolimits _{i=1}^{N}(c_i-{\bar{c}}_N)^2\) and \(\Delta _N=\max \limits _{1\le i \le N}c_i-\min \limits _{1\le i \le N}c_i\).
Proof
See Theorem 2.14.19 in [53]. \(\square \)
Proof of Lemma 1
Proof
For \({\varvec{p}}^*\), the optimal solution of (SP), we have
where inequality (a) is due to that if otherwise, \({\varvec{p}}^*\) cannot be the optimal solution because it will give a larger objective value of \(f({\varvec{p}})\) than setting \({\varvec{p}}={\varvec{0}}.\) Given the non-negativeness of \({\varvec{p}}^*\), we have \(\Vert {\varvec{p}}^*\Vert _2 \le \Vert {\varvec{p}}^*\Vert _1.\) So we obtain the first inequality in the lemma.
For \({\varvec{p}}_t\) specified by Algorithm 1, we have,
where the first inequality comes from the projection (into the non-negative orthant) step in the algorithm. Note that
Therefore,
and it holds with probability 1.
Next, we establish that when \(\Vert {\varvec{p}}_{t}\Vert _2\) is large enough, then it must hold that \(\Vert {\varvec{p}}_{t+1}\Vert _2 \le \Vert {\varvec{p}}_{t}\Vert _2\). Specifically, when \(\Vert {\varvec{p}}_{t}\Vert _2\ge \frac{2{\bar{r}}+m({\bar{a}}+{\bar{d}})^2}{{\underline{d}}}\), we have
when \(\gamma _t\le 1.\) On the other hand, when \(\Vert {\varvec{p}}_{t}\Vert _2\le \frac{2{\bar{r}}+m({\bar{a}}+{\bar{d}})^2}{{\underline{d}}}\),
where (b) comes from the triangle inequality of the norm.
Combining these two scenarios with the fact that \({\varvec{p}} _1 = {\varvec{0}}\), we have
for \(t=1,...,n\) with probability 1.
\(\square \)
Proof of Theorem 1
Proof
First, the primal optimal objective value is no greater than the dual objective with \({\varvec{p}}={\varvec{p}}^*.\) Specifically,
Taking expectation on both sides,
Thus, the optimal objective value of the stochastic program (by a factor of n) is an upper bound for the expected value of the primal optimal objective. Hence
where the expectation is taken with respect to \((r_t, {\varvec{a}}_t)\)’s. In above, the second line comes from the optimality of \({\varvec{p}}^*\), while the third line is valid because of the independence between \({\varvec{p}}_t\) and \((r_t, {\varvec{a}}_t).\)
The importance of the above inequality lies in that it relates and represents the primal optimality gap in the dual prices \({\varvec{p}}_t\) – which is the core of Algorithm 1. From the updating formula in Algorithm 1, we know
Moving the cross-term to the right-hand side, we have
Consequently,
hold for all n and distribution \({\mathcal {P}}\in \Xi .\)
For the constraint violation, if we revisit the updating formula, we have
where the inequality is elementwise. Therefore,
In the second line, we remove the term involve \({\varvec{p}} _1\) with the algorithm specifying \({\varvec{p}} _1={\varvec{0}}.\) Then with Lemma 1, we have
\(\square \)
Proof of Proposition 2
Proof
Define SLP\((s, {\varvec{b}}_0)\) as the following LP
where \({\varvec{b}}_0 = (b_{01},...,b_{0m})\) denotes the constraint relaxation quantity for the scaled LP. Denote the optimal objective value of SLP\((s, {\varvec{b}}_0)\) as \(R^*(s, {\varvec{b_0}}).\) Also, denote \({\varvec{x}}({\varvec{p}}) = (x_1({\varvec{p}}),...,x_n({\varvec{p}}))\) and \(x_j({\varvec{p}})=I(r_j>{\varvec{a}}_j^\top {\varvec{p}}).\) It denotes the decision variables we obtain with a dual price \({\varvec{p}}.\)
We prove the following three results:
-
(i)
The following bounds hold for \(R_n^*,\)
$$\begin{aligned} \sum _{j=1}^n r_jx_j({\varvec{p}}_n^*)-m{\bar{r}}+ \le R_n^* \le \sum _{j=1}^n r_jx_j({\varvec{p}}_n^*) + m{\bar{r}}. \end{aligned}$$ -
(ii)
When \(s\ge \max \{16{\bar{a}}^2,\exp {\{16{\bar{a}}^2\}},e\},\) then the \({\varvec{x}}({\varvec{p}}_n^*)\) is a feasible solution to SLP\(\left( s, \left( \sqrt{s}\log s+\frac{ms}{n}\right) {\varvec{1}}\right) \) with probability no less than \(1-\frac{m}{s}\).
-
(iii)
The following inequality holds for the optimal objective values to the scaled LP and its relaxation
$$\begin{aligned} R_s^* \ge R^*\left( s, \left( \sqrt{s}\log s+\frac{ms}{n}\right) {\varvec{1}} \right) - \frac{{\bar{r}}\sqrt{s}\log s }{{\underline{d}}}-\frac{{\bar{r}}sm}{{\underline{d}}n}. \end{aligned}$$
For part (i), this inequality replace the optimal value with bounds containing the objective values obtained by adopting optimal dual solution. The left-hand side of the inequality comes from the complementarity condition while the right-hand side can be shown from Lemma 2.
For part (ii), the motivation to introduce a relaxed form of the scaled LP is to ensure the feasibility of \({\varvec{p}}_n^*\). The key idea for the proof is to utilize the feasibility of \({\varvec{p}}_n^*\) for (P-LP). To see that, let \(\alpha _{ij}=a_{ij}I(r_j>{\varvec{a}}_j^T{\varvec{p}}^*)\) and
Here the first and third inequality comes from the bounds on \(a_{ij}\)’s, and the second one comes from the feasibility of the optimal solution for (P-LP).
Then, when \(k>\max \{16{\bar{a}}^2,\exp {\{16{\bar{a}}^2\}},e\}\), by applying Hoeffding-Bernstein’s Inequality
for \(i=1,...,m\). Here inequality (e) comes from (2), (f) comes from applying Lemma 5, and (g) holds when \(k>\max \{16{\bar{a}}^2,\exp {\{16{\bar{a}}^2\}},e\}\).
Let event
and \(E=\bigcap \nolimits _{i=1}^{m}E_i\). The above derivation denotes the case where the i-th constraint is satisfied for the LP with constraint relaxation. By applying union bound, we obtain \({\mathbb {P}}(E)\ge 1-\frac{m}{s}\) and it completes the proof of part (ii).
For part(iii), denote the dual optimal solution to SLP\(\left( s, \left( \sqrt{s}\log s+\frac{ms}{n}\right) {\varvec{1}}\right) \) as \(\tilde{{\varvec{p}}}_s.\)
where the first inequality comes from dual optimality of \(\tilde{{\varvec{p}}}_s^*\) and the second inequality comes from the upper bound of \(\Vert {\varvec{p}}_s^*\Vert \) and the duality of the scaled LP \(R_s^*\). Therefore,
Finally, we complete the proof with the help of the above three results.
where \({\mathbb {I}}_E\) denotes an indicator function for event E. The first line comes from applying part (iii) while the second line comes from the feasibility of \({\varvec{p}}^*\) on event E. Then,
where the first line comes from part (ii) – the probability bound on event E, the second line comes from the symmetry of the random permutation probability space, and the third line comes from part (i). We complete the proof. \(\square \)
Proof of Theorem 3
Proof
For the regret bound,
where \(x_t\)’s are specified according to Algorithm 1. Then
where \({\tilde{R}}_{n-t+1}^*\) is defined as the optimal value of the following LP
For the first part of (3), we can apply Proposition 2. Meanwhile, the analysis of the second part takes a similar form as (1) in the analysis of the previous stochastic input model. Specifically,
Similar to the stochastic input model,
Thus, we have
Combine two parts above, finally we have
Thus, we complete the proof for the regret. The proof for the constraint violation part follows exactly the same way as the stochastic input model. \(\square \)
Proof for Theorem 4
Proof
The proof follows mostly the proof of Theorem 1 and Theorem 3. We only highlight the difference here. First, the sample average approximation form of the dual problem takes a slightly different form but it is still convex in \({\varvec{p}}.\)
The updating formula for \({\varvec{p}}_t\) is different but we can achieve the same relation between \({\varvec{p}}_t\) and \({\varvec{p}}_{t+1}.\) At time t, if \(\max \limits _{l=1,...,k}\left\{ r_{jl}-{\varvec{a}}_{jl}^\top {\varvec{p}}\right\} >0\), we have
while if \(\max \limits _{l=1,...,k}\left\{ r_{jl}-{\varvec{a}}_{jl}^\top {\varvec{p}}\right\} \le 0\), we have
Combining those two parts, we obtain
which is the same formula as the one-dimensional setting. With the above results, the rest of the proof simply follows the same approach as the one-dimensional case.
\(\square \)
Proofs for results in Section 6
1.1 Proof for Lemma 3
Proof
We have
For the first line, on the right-hand side, the two inner expectations are taken with respect to a uniform random sampling on \({\varvec{Z}}_t\) and \(\tilde{{\varvec{Z}}}_{t'}\) respectively. Specifically, \({\varvec{Z}}_{t-s}\) (or \(\tilde{{\varvec{Z}}}_{s}\)) can be viewed as a random sampled subset from \({\varvec{Z}}_t\) (or \(\tilde{{\varvec{Z}}}_{t'}\)). For the second line, the first part comes from Jensen’s inequality and the expectation in the second part is taken with respect to the random sampling of \({\varvec{Z}}_t\) from \({\varvec{Z}}_n.\) \(\square \)
1.2 Proof for Lemma 4
Proof
For the first inequality, we refer to Theorem 3 in [44]. For the second inequality, it is a direct application of Massart’s Lemma (See Lemma 26.8 of [54]). \(\square \)
1.3 Proof for Proposition 5
Proof
Let \({\mathcal {F}}_{{\varvec{p}}} = \left\{ f_{{\varvec{p}}}: f_{{\varvec{p}}}(r,{\varvec{a}}) = rI\left( r>{\varvec{a}}^T{\varvec{p}} \right) \right\} \), \({\varvec{Z}}_t = \{(r_1,{\varvec{a}}_1),...,(r_t,{\varvec{a}}_t)\},\) and \({\tilde{Z}}_{n-t} = \{(r_{t+1},{\varvec{a}}_{t+1}),...,(r_n,{\varvec{a}}_n)\}.\) Also, we assume \(n-t>t\) without loss of generality. Then,
Here the first line comes from taking maximum over \({\mathcal {F}}_{{\varvec{p}}}\), the second line comes from Lemma 3 and the third line comes from lemma 4.
Similarly, we can show that the inequality on \(a_{ij}\)’s holds. Thus the proof is completed. \(\square \)
1.4 Proof for Theorem 6
Proof
At time \(t+1,\)
where the expectation is taken with respect to the random permutation. The first line comes from the algorithm design, the second line comes from the symmetry over the last \(n-t\) terms, and the last line comes from the application of Proposition 5. To relate the first term in the last line with the offline optimal \(R_n^*,\) we utilize Proposition 2. Then the optimality gap of Algorithm 3 is as follows,
where the third line comes from that \(r_1x_1\ge 0\) because \(x_1=1\) if and only if \(r_1>0\), and the last line comes from an application of Proposition 2. Next, we analyze the constraint; again, from Proposition 5, we know
where the first inequality comes from Proposition 5, the second inequality comes from the feasibility of the boundedness assumption and the scaled LP solved at time t, and the last inequality comes from the general position assumption that there are at most m columns such that \(r_j={\varvec{a}}_j^\top {\varvec{p}}_{t}\) for \(j=1,...,n\) and any fixed \(t=1,...,T\). Due to the symmetry of the random permutation,
Summing up the inequality, we have
which implies
\(\square \)
Proof of Theorem 7
Proof
At time t, the optimal solution to the scaled LP is \(\tilde{{\varvec{x}}}^{(t)}=({\tilde{x}}_1^{(t)},...,{\tilde{x}}_t^{(t)})\). We have
Then, for the objective,
where the second line comes from an application of Proposition 2. Then, we analyze the constraint violation. From the construction of the algorithm, we have that \({\varvec{E}}[a_{it}x_t]\le d_i\). Let
and then we know
is a supermartingale with \(\vert A_{ij}\vert \le {\bar{a}}+{\bar{d}}\). Then if we apply Hoeffding’s lemma for supermartingale, we have
when \(n>3\). Thus,
\(\square \)
Proof of Theorem 8
Proof
For the upper bound of v, lemma 1 and
can give the result.
For the fesibility, first, we prove that if the initial solution is infeasible for the i-th constraint, i.e., \(\sum \nolimits _{t=1}^{n} a_{it}x_{t}>nd_i\), then
with high probability. According to the definition of v, (4) says the corresponding \({\hat{x}}_t\) is feasible. To start, (4) holds with probability 1 if \(\vert S_0\vert =n_+\) which implies \(S_0=S_+\). Otherwise, the infeasibility indicates \(n_+\ge \max \{1,frac{{\underline{d}}n}{{\bar{a}}}\}\), and by definition \(n_+ \le n.\)
This implies bounds on the cardinality of \(S_0\),
and
Consequently,
Then, since \(n\ge \left( \frac{6{\bar{a}}}{{\underline{d}}}\right) ^4\)
Moreover, for any i s.t. \(\sum \nolimits _{t=1}^n a_{it} < \frac{n{\underline{d}}}{2}\), since \(n>16\) and \(\sqrt{n}>\frac{12{\bar{a}}({\bar{r}}+({\bar{a}}+{\underline{d}})^2m)\log n}{{\underline{d}}^2}\), we have
Then, for any i s.t. \(\frac{n{\underline{d}}}{2}\le \sum \nolimits _{t=1}^{n}a_{it}\le {n{\underline{d}}}\). Similarly, we can find that
Combining three parts above, we have that with probability at least \(1-\frac{2m}{n^2}\), the modified solution is feasible. Given the fact that the modified solution change the original solution for at most \(O(\sqrt{n}\log n)\) entries, the modified solution can achieve \(O((m+\log n)\sqrt{n})\) regret. \(\square \)
Rights and permissions
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Li, X., Sun, C. & Ye, Y. Simple and fast algorithm for binary integer and online linear programming. Math. Program. 200, 831–875 (2023). https://doi.org/10.1007/s10107-022-01880-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10107-022-01880-x