Karen Elliott

Existing black-box portfolio management systems are prevalent in the financial industry due to commercial and safety constraints, although their performance can fluctuate dramatically with changing market regimes. Evaluating these non-transparent systems is computationally expensive, as fixed budgets limit the number of possible observations. Therefore, achieving stable and sample-efficient optimization for these systems has become a critical challenge. This work presents a novel Bayesian optimization framework (TPE-AS) that improves search stability and efficiency for black-box portfolio models under these limited observation budgets. Standard Bayesian optimization, which solely maximizes expected return, can yield erratic search trajectories and misalign the surrogate model with the true objective, thereby wasting the limited evaluation budget. To mitigate these issues, we propose a weighted Lagrangian estimator that leverages an adaptive schedule and importance sampling. This estimator dynamically balances exploration and exploitation by incorporating both maximization of model performance and minimization of the variance of model observations. It guides the search from broad, performance-seeking exploration towards stable and desirable regions as the optimization progresses. Extensive experiments and ablation studies, which establish our proposed method as the primary approach and other configurations as baselines, demonstrate its effectiveness across four backtest settings with three distinct black-box portfolio management models.

In this chapter, we discuss the body of the turtle and provide a step-by-step guide to conducting a literature review in a systematic manner. In addition, we define which primary information sources can be accessed, revisit the role of theory in the literature review and include practical points on referencing and avoiding plagiarism.

In this chapter the ‘turtle template’ is introduced. This template organizes key information that needs to be considered when engaged in a research project in a structured and logical way.

This chapter looks back at the entire book to come up with 10 guiding principles or key take-aways that every researcher should take into account. They help researchers to stay on track, to keep the pace and to work towards an effective outcome.

In this chapter the book’s context (the rigor/relevance debate in business schools) and the target audience (students/executive program participants in business schools or economic departments in universities) are introduced. The structure of the book, mirroring the structure of a research project, is outlined, and key components (bias, elaboration and practitioner boxes are introduced.

In this chapter we introduce the 5C framework as the first leg of the turtle. In this leg the raison d’etre and overall approach of the proposed research project is justified by outlining it’s common ground, complication, concern, course of action, and expected contribution.

In this chapter, we discuss preserving the ethical integrity of research. We introduce ethics, the implications for research, compliance and safeguarding the research participants, context, and the researcher themselves. Methods for evaluating the likely ethical issues during the research process are provided and we discuss ethical governance in both physical and online environments.

This chapter presents the most common business and management data collection strategies, and methods for quantitative and qualitative research. We include survey design and conducting interviews, focus groups or observation. We define the primary differences between choosing quantitative or qualitative data collection methods, highlight common mistakes and present resolutions in preserving rigor. The chapter concludes by documenting the other types of data collection methods on offer for business and management research.

In this chapter we focus on the operationalization of the variables used in the research question and conceptual model. We suggest the use of an operationalization table that breaks down concepts in dimensions, indicators and suggested measurements for each indicator.