Overview

A prime-brokerage division of a large financial firm was struggling to get the optimal pairings of investment managers with the investors for their annual investor conference meetings. The Optimal Solutions Inc (OSI) team was hired to provide a mathematical model which optimally assigns investor – investment manager pairings for all available time-slots. The objective was to arrive at the conference schedule with the best overall satisfaction ratings from the attendees. 

‍

The OSI team was able to provide a platform to create an end-user application that could consistently and accurately generate an optimal conference schedule based on registration data. The scheduling application developed by Optimal Solutions helped the company to maximize the numbers of investor meeting requests fulfilled.

‍

Business Description and Background

A finance company's brokerage unit holds several conferences a year where fund managers and investors meet and discuss investment opportunities. Prior to the event, investors register and submit a list of managers they would most like to meet as well as time-slots they are available for meetings. While one-on-one meetings are most desirable, resource constraints require that some fund managers must hold meetings where several investors participate at once (group meetings). It is the objective of the organizer to create a conference schedule that will maximize the number of investor meeting requests fulfilled while minimizing the frequency and size of group meetings (subject to business constraints). Also, the ultimate goal is to satisfy the fund managers' as well as the investors' expectations under the logistics and time constraints. Even for a modest number of attendees, due to the explosion of the number of possibilities, this problem becomes very complex for routine approaches using tools like Excel. 

‍

The Challenge

Optimal Solutions was charged with the task of delivering an Optimization Tool that: 

• Optimally creates alternate schedule using same data;
• Provides robust data interchange capabilities so as to allow frequent data updates that reflect the changing reality;
• Provide a system which prevents double bookings;
• Provide statistical analysis to objectively generate reports;
• Provides a flexible and user-friendly interface.

‍

The Solution

Optimal Solutions’ team utilized the AIMMS development platform to create an end-user application that could repeatably generate an optimal conference schedule based on registration data downloaded from the web. The underlying mathematical model solved a mixed-integer program to optimally assign investor/manager pairings for all available time-slots. The model addressed several business constraints using both pre-solve data configuration as well as live run-time elements. This featured the use of Mixed Integer Programming (MIP). 

‍

Key Results & Benefits

Our tool provides the following benefits – 

‍

1. Time to generate a single candidate conference schedule reduced from twenty-plus hours to 20 minutes;
2. We achieved a 99% feasibility for investor requests successfully scheduled;
3. Organizer now has the control to generate schedules with an objective in mind;
4. Organizer’s focus has shifted from, “Let’s try and get a schedule finished on time,” to, “Let’s create 10 different schedules using slightly different settings and decide which one we like best”;
5. Much improved error-catching system that prevents double-bookings and time-conflicts;
6. Organizer has access to valuable post-solve printable reports that provide statistics and distributions related to the generated schedule;
7. The application effectively reduces the amount of business knowledge required by an employee to build a good schedule;

‍