An overview of MOSEK Erling D. Andersen MOSEK ApS Fruebjergvej 3, - PowerPoint PPT Presentation

An overview of MOSEK Erling D. Andersen MOSEK ApS Fruebjergvej 3, Symbion Science park, Box 16 2100 Copenhagen O Denmark Email: sales@mosek.com WWW: http://www.mosek.com http://www.mosek.com October 5, 2013

Introduction 2 / 25

Overview Introduction 1. Background. Overview 2. Product overview. Background 3. Customers. Product overview 4. Product details. Customers 5. Licensing. Product details 6. Getting started. Getting started 7. The future. Licensing 8. Summary. Present and future of MOSEK Summary 3 / 25

Background 4 / 25

The company Introduction MOSEK ApS is a Danish company. ■ Background Vision: Create and sell software for solving linear and ■ The company convex optimization problems. Product overview Located in Copenhagen at Symbion Science Park. ■ Customers Daily management: Erling D. Andersen. ■ Product details Currently 4 full-time, 4 part-time employees and 1 phd ■ Getting started student. Licensing Present and future of MOSEK Summary 5 / 25

Product overview 6 / 25

MOSEK optimization tools Introduction Main product: MOSEK optimization tools. ■ Background Solves generic: ■ Product overview ◆ MOSEK Linear, quadratic, and nonlinear problems. Only optimization tools convex cases. Customers ◆ Conic optimization problems. Product details Getting started c T x min Licensing st. Ax b, = Present and future of MOSEK x ∈ K Summary where K is a convex cone (only linear, quadratic, semidefinite). ◆ Integer optimization problems. Same as above but some variables are integer ■ constrained. The software is NOT application specific. ■ 7 / 25

Customers 8 / 25

Types of customers Introduction Structural design. ■ Background ◆ Arch Bridge Analysis. Product overview Financial industry. Customers ■ Types of customers ◆ Portfolio optimization. Product details ◆ Trading and cash optimization. Getting started Licensing Power industry. ■ Present and future Forestry. ■ of MOSEK Bus crew scheduling. ■ Summary Analogue circuit design. ■ Bio tech. ■ Advertisement. ■ + more. ■ 9 / 25

Introduction End customer size. ■ Background ◆ Small with 1 user. Product overview ◆ Large with users e.g. large banks. Customers Types of customers OEM. ■ Product details ◆ Integrate MOSEK into a another product. Getting started Licensing Value added resellers. ■ Present and future of MOSEK Geographical markets: Summary North America. ■ Europa (few Danish customers). ■ Japan. ■ Australia. ■ South America. ■ 10 / 25

Product details 11 / 25

Optimizers Introduction Primal and dual simplex for linear problems. ■ Background Primal network simplex for network flow problems. ■ Product overview Interior-point for linear, conic and nonlinear problems. ■ Customers Two branch and bound and cut optimizers. ■ Product details Optimizers ◆ Freely included. Tuned for conic problems. Interfaces ◆ Paid addon: Tuned for linear problems. Getting started Licensing Present and future of MOSEK Summary 12 / 25

Interfaces Introduction Optimizer API ■ Background ◆ Matrix orientated. Product overview ◆ C, Java, .NET and Python. Customers Product details Fusion ■ Optimizers Interfaces ◆ Object orientated. Getting started ◆ Only linear and conic problems. Licensing ◆ Java, MATLAB, .NET and Python. Present and future of MOSEK Other ■ Summary ◆ AMPL ◆ MATLAB toolbox. ◆ R package. Third party links: ■ ◆ AIMMS, CVX, GAMS, MPL (soon), Woodstock. ◆ Coin OSI, Raven Toolbox, Yalmip, ... 13 / 25

Getting started 14 / 25

Useful links Introduction Software download: ■ Background http://mosek.com/download/ ◆ Product overview Trial license: Customers ■ Product details http://mosek.com/resources/trial/ ◆ Getting started Useful links Documentation: ■ Licensing http://docs.mosek.com ◆ Present and future of MOSEK Support: ■ Summary http://mosek.com/support/ ◆ Social medias: ■ ◆ http://mosek.com/resources/social-media/ 15 / 25

Licensing 16 / 25

Floating licenses Introduction PTS: Linear and quadratic optimization + integer ■ Background variables. Product overview PTON: Nonlinear and conic extension + integer variables. ■ Customers Product details PTOM: Mixed-integer optimizer extension. ■ Getting started One usage at any computer in the network. Token server ■ Licensing required. Floating licenses Server licenses Pricing: ■ Group license Academic license http://mosek.com/sales/pricing/ ◆ Present and future of MOSEK Summary 17 / 25

Server licenses Introduction PTS-NODE: Linear and quadratic optimization + integer ■ Background variables. Product overview PTON-NODE: Nonlinear and conic extension + integer ■ Customers variables. Product details PTOM-NODE: Mixed-integer optimizer + integer ■ Getting started variables. Licensing No token server. ■ Floating licenses Server licenses Unlimited use on a single prespecified computer. ■ Group license Pricing: ■ Academic license Present and future http://mosek.com/sales/pricing/ ◆ of MOSEK Summary 18 / 25

Group license Introduction A rental license. ■ Background Unlimited use of 1 or more features. ■ Product overview One organizational unit e.g. one department. ■ Customers No license token server. ■ Product details Price negotiated every third year. ■ Getting started Licensing Floating licenses Server licenses Group license Academic license Present and future of MOSEK Summary 19 / 25

Academic license Introduction Free for both personal and classroom usage. ■ Visit: Background ■ Product overview ◆ Customers http://mosek.com/resources/academic-license/ Product details Getting started Licensing Floating licenses Server licenses Group license Academic license Present and future of MOSEK Summary 20 / 25

Present and future of MOSEK 21 / 25

Present Introduction MOSEK version 7 released sping 2013. ■ Background ◆ Added support semidefinite optimization. Product overview ◆ Added the linear and conic Fusion modeling API. Customers ◆ Added a new mixed integer optimizer. Product details Getting started Strong emphasize on linear and conic problems. ■ Licensing Present and future of MOSEK Present The future Summary 22 / 25

The future Introduction Continued strong emphasize on linear and conic ■ Background optimization + integer variables. Product overview Emphasize improvements: ■ Customers ◆ Improve speed and stability. Product details ◆ Exploit hardware better. E.g. AVX and Intel Getting started Licensing New features: ■ Present and future of MOSEK ◆ Add mixed-integer semidefinite optimization. Present ◆ Add nonsymmetric cones (if algorithms deemed good The future enough). Summary 23 / 25

Summary 24 / 25

Final comments Introduction Presented MOSEK. ■ Background Slides are available at: ■ Product overview http://www.mosek.com/presentations/ ◆ Customers Product details Getting started Licensing Present and future of MOSEK Summary Final comments 25 / 25