TFO -- Hiring and Outsourcing in the Online Labor market

1. Algorithms for Hiring and Outsourcing in the
Online Labor Market
Presenter: Maggie Zhang

2. Team Formation with Outsourcing
(Model – TFO)
Problem – outsourcing and hiring workers in a general setting, where
workers form a team and contribute different skills to perform a task.
Assigned w/t Extra
Assigned w Extra
Hired
/Fired
w/t Priori
Minimize Labor Costs
Hired Worker – Fixed Salary
with Hiring/Firing Cost
Outsourcing – Extra
Premium

3. TFO Model Summary
• Objectives: Load balancing & Compatibility
• Constraints: Worker Capacity
• Algorithm Setup: Online or Offline
(1) if the cost of hiring or firing workers is too high,
outsourcing becomes preferable to hiring;
(2) if the cost of outsourcing work relative to salaries
of hired workers is too high, hiring becomes
preferable to outsourcing;
(3) if the workload consists of many repetitions of
similar tasks, hiring becomes preferable to
outsourcing.

4. Characteristics
(1) at each step, an entirely new
instance of the set-cover problem
needs to be solved by using hired and
outsourced workers;
(2) hired and outsourced workers
collaborate with each other, and this
needs to be taken into account;
(3) workers can be hired, fired, hired
again, and so on, so one has to keep
track of their status at every point in
time.
1 T*
1 n
t

5. Notation

6. Problem Definition
Competitive Ration (CR)

7. Background Problems
Goal: Find the team to cover all skills
and minimize outsourcing expenses.
Goal: when , hire the
worker and achieve CR of 2.

8. LUMPSUM
Statement – Combination of SetCover and SkiRental
New Notation –
: the set of skills of that cannot be covered by already-hired workers.
: the accumulated costs.
Arbitrarily bad CR - M

9. LumpSum LP

10. LumpSum Algorithm

11. TFO Problem
Statement – add salary and fire cost to Lumpsum.
New Notation –
: intervals in which workers are hired
Good CR

12. ALT-TFO Algorithm

13. Experiments
Methods: TFO/Lumpsum, TFO/Lumpsum-Heuristic,
Always-Outsource, Always-Hire, TFO-Adaptive

14. Total Cost
over Tasks

15. Total Cost
over Tasks

16. Related Work
CrowdSourcing: MAX of workers’ skill match
Team Formation: MIN of team’s communication cost
Cache Usage: Page inserted.

17. Conclusion
• Online Primal-Dual Technique
• Extended to: Limit the
number of workers.