father and son duo take on global logistics with optimal dynamics’ sequential decision ai platform
The Evolution of AI and Stochastic Optimization
Terms like “innovation,” machine learning, and artificial intelligence are frequently used, yet often lack specific definition regarding their actual implications. AI/ML encompasses a vast spectrum of research areas, addressing diverse problems through varied, and sometimes conflicting, solution methodologies.
A well-established research domain, with roots extending back to the mid-20th century, centers around stochastic optimization – the process of making decisions amidst uncertainty to achieve a desired outcome. A typical example involves optimizing an airline’s flight schedule to maximize profitability. Airlines must finalize schedules months in advance, despite unpredictable factors like weather patterns and fluctuating route demand, or even unforeseen events like global pandemics impacting travel. This remains a dynamic field, fundamentally influencing much of contemporary life.
Sequential Decision Analytics and Optimal Dynamics
For decades, Warren B. Powell has investigated this challenge as a researcher at Princeton University, leading the Castle Lab. His work focuses on unifying different areas of stochastic optimization under a single framework, termed “sequential decision analytics.” This framework is designed to optimize problems where each decision impacts subsequent choices.
Problems of this nature are prevalent in sectors such as logistics, scheduling, and other critical business functions. The Castle Lab has consistently collaborated with industry partners and secured substantial grant funding, totaling tens of millions of dollars, throughout its history.
After years of research, Powell partnered with his son, Daniel Powell, to transform his accumulated research and expertise into a commercially viable startup, Optimal Dynamics. Warren Powell has since transitioned to a full-time role as chief analytics officer, while Daniel Powell assumed the position of CEO.
Recent Funding and Growth
Last week, the company secured $18.4 million in new funding, led by Mike Droesch of Bessemer, who was recently promoted to partner within the firm’s $3.3 billion fund. Currently, Optimal Dynamics employs 25 individuals and is headquartered in New York City.
CEO Powell explained that the journey to establishing the company, beginning in mid-2017 with a $450,000 pre-seed round, involved a period of exploration to identify a suitable product-market fit. He acknowledged that “getting this technology right is not easy.”
Focus on the Trucking Industry
The company ultimately concentrated on the trucking industry, recognizing its inherent complexity and suitability for sequential decision-making – a field in which the elder Powell has dedicated his career. According to the CEO, “Within truckload, there’s a series of uncertain variables.” Optimal Dynamics positions itself as “the first company that can learn and plan for an uncertain future.”
Recent years have witnessed significant venture capital investment in logistics and trucking, driven by the potential for substantial disruption in this large and fragmented market. However, rather than creating a new marketplace or pursuing a vertically integrated approach, Optimal Dynamics opted for a simpler enterprise SaaS model, aiming to enhance optimization for existing businesses.
Demonstrated Results and Cost Savings
An early customer, operating a fleet of 120 power units, reportedly saved $4 million by utilizing the company’s software. This was achieved through improved equipment utilization and streamlined operations. The customer was able to “sell off about 20 vehicles that they didn’t need anymore due to the underlying efficiency,” Powell stated.
Furthermore, the company reduced a team of 10 logistics managers to a single individual responsible for handling exceptions. As an illustration, Powell cited a scenario where a driver unexpectedly resigned mid-route, leaving a shipment stranded. “Trying to train a computer on weird edge events [like that] is hard,” he noted.
Optimal Dynamics delivers cost savings to customers primarily through enhanced equipment utilization, with employee cost reductions through automation playing a secondary role. Powell emphasized that utilization typically offers the greatest potential for impact.
Technical Innovation: Value Function Approximation
A key technical advancement developed by the company is the ability to efficiently solve the highly complex optimization problems faced by logistics companies. This is accomplished through value function approximation, a technique that estimates the outcomes of decisions rather than exhaustively calculating all possible stochastic optimization solutions, thereby reducing computational demands.
Powell explained that the company excels at “taking in this extraordinary amount of detail while handling it in a computationally efficient way,” establishing a unique position within the market.
Funding and Future Expansion
Positive early customer results led to a $4 million seed round led by Homan Yuen of Fusion Fund, a firm specializing in technically advanced startups. Powell described the fundraising process as challenging, coinciding with the initial stages of the pandemic last year. A corporate fund withdrew its commitment at the last moment, creating a period of uncertainty. However, the subsequent Series A round was significantly smoother, closing in just 17 days.
With the new funding secured, Optimal Dynamics plans to expand its workforce from 25 to 75 employees throughout the year, with personnel gradually returning to the company’s office in Manhattan’s Flatiron neighborhood. The company targets customers operating fleets of 75 trucks or more, including both rental fleets and private fleets owned by companies like Walmart.